DEVELOPMENT AND APPLICATION OF NEW APPROACHES FOR STUDIES OF INFLUENZA INDUCED INFLAMMATION AND DNA DAMAGE

... Investigation of DNA damage as a potential mechanism of influenza pathogenesis 66 3.1.2 Phosphorylation of H2AX (Ser-139) during DNA strand breaks 69 3.1.3 DNA damage is triggered by influenza. .. to sense for DNA damages, activate cell signaling and repair DNA lesions DNA damage are surveyed via DNA damage sensors, Mre11-Rad50-Nbs1 (MRN) complex, and replication protein A (RPA) and Rad9-Rad1-Hus1... more harmful DNA strand breaks Taken together, DNA damage (in the form of DNA strand breaks) can be induced directly by ROS / RNS, as well as indirectly during the repair of ROS / RNS -induced base

DEVELOPMENT AND APPLICATION OF NEW APPROACHES FOR STUDIES OF INFLUENZA-INDUCED INFLAMMATION AND DNA DAMAGE Li Na B.Sc. (Hons), NUS A THESIS SUBMITTED FOR THE DEGREE OF DOCTOR OF PHILOSOPHY IN SCIENCE DEPARTMENT OF MICROBIOLOGY YONG LOO LIN SCHOOL OF MEDICINE NATIONAL UNIVERSITY OF SINGAPORE 2015 1 i. Declaration: I hereby declare that the thesis is my original work and it has been written by me in its entirety. I have duly acknowledged all the sources of information which have been used in the thesis. This thesis has not been submitted for any degree in any university previously. _______________________ Li Na i ii. Acknowledgements: First and foremost, I would like to express my sincere gratitude to my supervisors Professor Bevin Engelward and Professor Vincent Chow, for offering me the opportunity to be a part of their labs, for their intellectual and moral support, for their inspiration in doing good science, and for teaching me how to present myself and my work. I would like to show my appreciation to my thesis advisory committee members Professor Fred Wong and Professor Marie Clement for their invaluable comments and suggestions for my thesis, and to my collaborators, Dr Damien Thévenin and Professor Donald Engelman. Many thanks to my lab mates and colleagues in Singapore – Dr Yamada Yoshiyuki, Prashant, Tze Khee, Kai Sen, Anandi and Dr Orsolya Kiraly- for lending me support in my projects, for giving me constructive suggestions, and most of all, for your friendship. I will like to specially mention my appreciation towards Dr Yamada, who had mentored me for 3 years and had taught me useful skills and good lab practices. In addition, my project would not have been possible if not for Mrs Phoon who painstakingly propagated all the viruses. To my lab mates in MIT- Marcus, Jenny, Shelly, Lizzie, Yang Su, Jing- thank you for being such wonderful colleagues and for the good scientific exchanges we have had. To all my SMART friends, the “Aunties of SMART”, the lung repair group, and my fellow SMA3 students, thank you for being great colleagues and amazing friends too, making my time in SMART an extremely enjoyable one. I am going to miss you guys! And most importantly, I would like to thank my parents for believing in me and for being my pillar of support. And last but not least, to my boyfriend, for being there with me, always, to listen, to share, to advice and to empathize. Thank you all for ii giving me the strength, courage and wisdom needed to complete this long and demanding journey. Some passages and images in chapter 2 and chapter 4 are quoted verbatim or reprinted from “Na Li, Yin Lu, Thevenin Damien, Yamada Yoshiyuki, Limmon Gino, et al. (2013) Peptide targeting and imaging of damaged lung tissue in influenzainfected mice. Future Microbiol 8: 257-269.” with permission from Future Microbiology. Some passages and images from chapter 2 and chapter 5 are quoted and reprinted from “Sukup-Jackson R. Michelle, Kiraly Orsolya, Kay Jennifer, Na Li, Rowland, Kelly E. Winther Elizabeth A., Chow Danielle N., Kimoto Takafumi, Matsuguchi Tetsuya, Jonnalagadda Vidya S., Maklakova Vilena I., Singh Vijay R., Wadduwage Dushan N., Rajapakse Jagath, So Peter T. C., Collier Lara S., Engelward Bevin P. (2014) Rosa26-GFP direct repeat (RaDR-GFP) mice reveal tissue- and age-dependence of homologous recombination in mammals in vivo. PLoS Genet 10: e1004299.” PLoS Genetics is an open-access journal, so no permission from the publisher was needed for using the materials. Passages quoted in this thesis were contributed to the journal by the author of this thesis. iii iii. Funding: This thesis is supported by the Singapore National Research Foundation (NRF), Ministry of Education (MOE) and in part by the National Institute of Environmental Health Sciences (P01-ES006052). The views expressed herein are solely the responsibility of the author and do not necessarily represent the official views of the funding bodies. iv iv. Publications: Yoshiyuki Yamada, Gino V. Limmon, Dahai Zheng, Na Li, Liang Li, Lu Yin, Vincent T. K. Chow, Jianzhu Chen, Bevin P. Engelward. (2012) Major shifts in the spatiotemporal distribution of lung antioxidant enzymes during influenza pneumonia. PLoS One 7: e31494. Na Li, Lu Yin, Damien Thévenin, Yoshiyuki Yamada, Gino Limmon, Jianzhu Chen, Vincent TK Chow, Donald M Engelman, Bevin P Engelward. (2013) Peptide targeting and imaging of damaged lung tissue in influenza-infected mice. Future Microbiol 8: 257-269. Sukup-Jackson R. Michelle, Kiraly Orsolya, Kay Jennifer, Na Li, Rowland, Kelly E. Winther Elizabeth A., Chow Danielle N., Kimoto Takafumi, Matsuguchi Tetsuya, Jonnalagadda Vidya S., Maklakova Vilena I., Singh Vijay R., Wadduwage Dushan N., Rajapakse Jagath, So Peter T. C., Collier Lara S., Engelward Bevin P. (2014) Rosa26-GFP direct repeat (RaDR-GFP) mice reveal tissue- and age-dependence of homologous recombination in mammals in vivo. PLoS Genet 10: e1004299. Na Li, Marcus Parrish, Tze Khee Chan, Lu Yin, Prashant Rai, Yamada Yoshiyuki, Nona Abolhassani, Kong Bing Tan, Orsolya Kiraly, Vincent T. K. Chow, Bevin P. Engelward. Influenza infection induces host DNA damage and dynamic DNA damage responses during tissue regeneration. Cell Mol Life Sci. 2015 Mar 26. [Epub ahead of print] v v. Conference Abstracts: Na Li, Liang Li, Damien Thévenin, Yoshiyuki Yamada, Yin Lu, Gino Limmon, Vincent TK Chow, Donald M Engelman, Bevin P Engelward. Peptide targeting and imaging of damaged lung tissue in influenza-infected mice. Poster session presented at: Asia- Pacific Congress of Medical Virology; 2012, 6-8 June; Adelaide, Australia. Na Li, Damien Thévenin, Liang Li, Yoshiyuki Yamada, Yin Lu, Gino Limmon, Vincent TK Chow, Donald M Engelman, Bevin P Engelward. Peptide targeting and imaging of damaged lung tissue in influenza-infected mice. Poster session presented at: Singapore - Japan Joint Forum On Emerging Concepts In Microbiology; 2011, 15-16 November; Singapore. vi vi. Abbreviations: 5’-dRP 53BP1 8-OH-G / 8-OH-dG ALI ARDS AEI/ AEII ATM AP BALF BER CCSP CI DBD DDR DNA DNA-PK DNA-PKcs dpi DSBs H&E H2O2 5’- deoxyribose phosphate p53 Binding Protein 1 8-hydroxyguanosine/ 8-hydroxy-deoxyguanosine Acute lung injury Acute respiratory distress syndrome Alveolar epithelial type I/II cells Ataxia telangiectasia mutated Apurinic / apyrimidinic Bronchoalveolar lavage fluid Base excision repair Club cell secretary protein Contrast index DNA binding domain DNA damage response Deoxyribonucleic acids DNA- dependent protein kinase DNA- dependent protein kinase catalytic subunit days post infection DNA double strand breaks Hematoxylin and eosin Hydrogen peroxide HA HOCl hpi HR HRP iNOS M1/ M2 MDCK cells NA NAD+ NHEJ NO NP NS1/ NS2 NSAIDs NU7441 O2−• Haemagglutinin Hypochlorous acid hours post infection Homologous recombination Horse radish peroxidase Inducible nitric oxide synthase Matrix1/2 Madin-Darby canine kidney cells Neuraminidase Nicotinamide adenine dinucleotide Non-homologous end joining Nitric oxide Nucleoprotein Non-structural protein 1/ Non-structural protein 2 Non-steriodal anti-inflammatory drugs 2-N-morpholino-8-dibenzothiophenyl-chromen-4-one Superoxide ONOOPA/ PB1/ PB2 PAR Peroxynitrite Polymerase A/ B1/ B2 Poly (ADP-ribose) vii PARP-1 PCNA Pdpn PEG PFU pHLIP PIKK RaDR-GFP PR8 Prdx6 Pro-SPC RdRp ROI ROS/RNS SOD TNF-α TUNEL XO Poly (ADP-ribose) polymerase 1 Proliferating cell nuclear antigen Podoplanin Polyethylene glycol Plaque forming units pH (Low) insertion peptide Phophatidylinositol-3-kinase-like kinases Rosa26 Direct Repeat-Green Fluorescent Protein Influenza A/Peurto Rico/8/34 Peroxiredoxin 6 Pro- surfactant protein C RNA-dependent RNA polymerase Region of interest Reactive oxygen and nitrogen species Superoxide dismutase Tumour necrosis factor-alpha Terminal deoxynucleotidyl transferase dUTP nick end labeling Xanthine oxidase viii TABLE OF CONTENTS I. DECLARATION: ........................................................................................... I II. ACKNOWLEDGEMENTS: ........................................................................... II III. FUNDING: ...................................................................................................IV IV. PUBLICATIONS: .........................................................................................V V. CONFERENCE ABSTRACTS: ...................................................................VI VI. ABBREVIATIONS: ....................................................................................VII TABLE OF CONTENTS .....................................................................................IX SUMMARY .......................................................................................................... 1 LIST OF FIGURES .............................................................................................. 4 LIST OF TABLES ............................................................................................... 7 CHAPTER 1 BACKGROUND ............................................................................. 8 1.1 Influenza virus ......................................................................................... 8 1.1.1 1.1.2 1.1.3 1.2 Influenza infection induces acute lung injury ..................................... 19 1.2.1 1.2.2 1.2.3 1.3 Cellular responses towards DNA damage ..............................................29 Base excision repair removes damaged bases ......................................30 DNA damage can arise from BER intermediates ....................................32 DNA double strand break (DSBs) repair .................................................33 Studying significance of DNA repair pathway in vivo ........................ 37 1.4.1 1.5 Influenza-induced cytopathy and lung injury .........................................20 Inflammatory responses and lung injury ................................................20 Reactive oxygen and nitrogen species during influenza infection........21 ROS and RNS can damage genomic DNA........................................... 26 1.3.1 1.3.2 1.3.3 1.3.4 1.4 Life cycle of Influenza virus.....................................................................10 Antigenic drift and antigenic shift ...........................................................14 Global burden imposed by influenza outbreaks.....................................17 Genetically engineered mouse that enable visualization of HR events 38 Improving biodistribution of therapeutic agents ................................ 39 1.5.1 Targeting inflamed tissue with pH sensitive peptide .............................39 ix 1.6 Thesis aims ........................................................................................... 41 CHAPTER 2 METHODS AND MATERIALS ..................................................... 42 2.1 Materials ................................................................................................ 42 2.1.1 2.1.2 2.1.3 2.1.4 2.1.5 2.1.6 2.2 Media, chemicals and reagents ...............................................................42 Cell cultures .............................................................................................43 List of antibodies for western blot and immunofluorescence ...............44 List of antibodies for flow cytometry ......................................................45 Source of influenza viruses .....................................................................45 pHLIP and pHLIP variant peptide sequences and conjugation .............46 Methods and protocols......................................................................... 47 2.2.1 Infection of mice and tissue collection ..................................................47 2.2.2 Lung homogenization and virus titration................................................48 2.2.3 Lung histology, infiltration index calculation and pathology analysis..48 2.2.4 Measurement of ROS markers and TNF-α ..............................................49 2.2.5 Western blotting.......................................................................................50 2.2.6 Flow cytometry of immune cells .............................................................50 2.2.7 Immunofluorescence assay ....................................................................51 2.2.8 Microscopy...............................................................................................52 2.2.9 Manual and semi-automated quantification of H2AX foci.....................53 2.2.10 Terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) and quantification ...................................................................................................54 2.2.11 Microarray data analysis .........................................................................55 2.2.12 NU7441 treatment of infected mice .........................................................56 2.2.13 RaDR mice and single cell suspension preparation ..............................56 2.2.14 RaDR cells RNA extraction and cDNA conversion .................................57 2.2.15 Direct PCR analysis using RNA transcripts............................................58 2.2.16 Nested PCR analysis for Full-length EGFP.............................................59 2.2.18 In vitro experiment with pHLIP ................................................................62 2.2.19 Peptide injection in mice .........................................................................62 2.2.20 Whole body and ex vivo whole organ bioimaging..................................62 2.2.21 Feature extraction and pHLIP quantification ..........................................63 2.2.22 Statistical analysis ...................................................................................65 CHAPTER 3 INFLUENZA INFECTION INDUCES DNA DAMAGE AND ROBUST DNA DAMAGE RESPONSES IN VIVO ............................................. 66 3.1 Introduction ........................................................................................... 66 3.1.1 Investigation of DNA damage as a potential mechanism of influenza pathogenesis ...........................................................................................................66 3.1.2 Phosphorylation of H2AX (Ser-139) during DNA strand breaks ............69 3.1.3 DNA damage is triggered by influenza infection in vitro .......................72 3.1.4 Aims of study ...........................................................................................74 3.2 Results................................................................................................... 76 3.2.1 Characterization of H1N1 murine model .................................................76 3.2.2 Prolonged inflammation after suppression of virus load.......................76 3.2.3 Broad characterization of pulmonary inflammation ...............................81 3.2.4 Oxidative stress is elevated during infection .........................................84 3.2.5 Host responses induce DNA damage in lung epithelium after influenza infection 86 3.2.6 DNA damage occurs in immune cell populations ..................................94 x 3.2.7 Influenza infection induces polymerization of poly (ADP- ribose) ........96 3.2.8 Influenza infection leads to apoptosis ....................................................99 3.2.9 Infection increases DNA damage among dividing cells.......................102 3.2.10 Influenza infection induces DSB repair proteins during tissue regeneration ..........................................................................................................105 3.2.11 NHEJ component, DNA-PK, is dispensable for DNA repair during tissue regeneration ..........................................................................................................109 3.3 Discussion........................................................................................... 113 3.4 Conclusion .......................................................................................... 120 3.5 Future studies ..................................................................................... 121 3.5.1 3.5.2 Verification of forms of DNA damage ...................................................121 DNA repair deficiencies on cell fate and disease outcome..................122 CHAPTER 4 FLUORESCENT DETECTION OF HOMOLOGOUS RECOMBINATION EVENTS FOLLOWING INFLUENZA INFECTION ........... 124 4.1 Introduction ............................................................................................. 124 4.1.1 4.1.2 4.1.3 4.2 Results and Discussion...................................................................... 131 4.2.1 4.2.2 4.2.3 4.2.4 4.3 Homologous recombination ..................................................................124 Approach for studying HR during influenza infection in vivo..............125 Aims of Study.........................................................................................129 PCR-based analysis of cDNA ................................................................131 Improved sensitivity with nested PCR ..................................................132 PCR analysis of HR mechanism in single cells ....................................136 EGFP expression to monitor HR events in lung epithelial cells ..........140 Conclusion and Future studies.......................................................... 143 CHAPTER 5 IMAGING AND TARGETING OF PH (LOW) INSERTION PEPTIDE (PHLIP) AT INFLAMED LUNG TISSUE DURING INFLUENZA INFECTION .. 144 5.1 Introduction ......................................................................................... 144 5.1.1 5.1.2 5.1.3 5.1.4 5.1.5 5.2 Targeted drug delivery as a therapeutic strategy.................................144 pHLIP: pH (low) insertion peptide .........................................................147 Inflammation changes pH of tissue microenvironment .......................151 Aims of study .........................................................................................152 Specific aims ..........................................................................................153 Results................................................................................................. 154 5.2.1 5.2.2 5.2.3 organs 5.2.4 5.2.5 5.2.6 5.2.7 5.2.8 pHLIP targets influenza-infected lungs.................................................154 pHLIP also targets other acidic tissue ..................................................156 pHLIP has high retention in inflamed lungs as compared to other 158 Design of pHLIP variants .......................................................................160 Evidence of pH-dependent pHLIP accumulation in infected lungs .....165 pHLIP accumulates in regions of lungs with heavy cellular infiltration 168 pHLIP colocalizes with damaged lung tissue .......................................172 pHLIP-targeted lung parenchyma has diminished antioxidant............175 xi 5.3 Discussion........................................................................................... 177 5.4 Conclusion .......................................................................................... 181 5.5 Future studies ..................................................................................... 181 5.5.1 pHLIP delivery of antioxidants ..............................................................181 CHAPTER 6 CONCLUSIONS ......................................................................... 183 REFERENCES: ............................................................................................... 188 xii Summary Influenza viruses are a group of highly contagious pathogens that account for significant morbidity and mortality worldwide. Extensive studies have suggested that dysregulated inflammatory responses including excessive production of reactive oxygen and nitrogen species (ROS / RNS) mediate lung injury in severe infections. However, underlying mechanisms and effective treatment strategies for inflammation-induced injury are not fully established. The aims of this thesis have been to broadly characterize an influenza-pneumonia mouse model, and to examine the impact of influenza infection on host genomes using the mouse model. Using the mouse model, the potential of a low pH targeting peptide as a delivery agent that can be used to improve treatment for inflammation-induced lung injury is also evaluated. One of the known mechanisms for ROS / RNS-induced toxicity is the ability to induce DNA lesions, which can lead to mutations and cell death. In the first study, we investigated whether DNA damage, a potential consequence of inflammationinduced ROS / RNS, is associated with severe influenza pneumonia. Using immunofluorescence techniques, we observed an increase in DNA damage, revealed by modified histones that form DNA repair foci at the sites of DNA strand breaks in the lung tissue of infected mice. DNA repair foci are induced in lung epithelial cells and infiltrating immune cells for prolonged duration even after viral clearance, suggesting that DNA damage is driven by inflammation. Notably, DNA damage was observed in lung tissue especially during the tissue regenerative phase of infection, when cell division occurs simultaneously with unresolved inflammation. If unrepaired, toxic DNA lesions, such as DNA double strand breaks (DSBs), can have dire consequences on cells, potentially leading to large scale 1 genetic rearrangements and cytotoxicity that can impede recovery process. Results show that DSB repair proteins Ku70 / Ku86 and Rad51 are upregulated during the later stage of infection, implicating DSB repair pathways in the resolution of DNA damage during lung tissue recovery. Taken together, the data demonstrate that DNA damage is associated with influenza pneumonia, and raise a possibility that DNA repair capacity may be a determinant for tissue recovery. While investigating disease mechanisms may set a foundation for identifying therapeutic targets for severe influenza infection, more effective treatment may be achieved by modifying delivery of existing treatments. Optimized drug delivery in terms of dose and biodistribution are important challenges for treating inflammation-induced lung injury during influenza infections. In the second part of this thesis, we investigated whether pH (Low) Insertion Peptide™ (pHLIP™), a peptide with high affinity for acidic microenvironments, can be used for specific delivery of drugs or imaging agents in influenza patients. This study is designed based on the understanding that inflamed tissue is acidic while healthy tissues are slightly alkaline. To test our hypothesis that pHLIP localize at sites of inflammation, fluorophore-conjugated pHLIP was injected into infected mice to track peptide distribution in vivo. Results show that pHLIP specifically targets inflamed lung during the later stage of infection, when severe pneumonia manifest. In addition, pHLIP-targeted lung tissues are injured and devoid of alveolar type I (AEI) and type II cells (AEII) that are found in healthy alveoli. Interestingly, pHLIP-targeted lung tissue is also characterized by depletion of Peroxiredoxin 6 (Prdx6), an enzymatic antioxidant abundantly expressed in the lung. Importantly, the specific distribution of pHLIP in inflamed tissue opens up opportunities for the delivery of pHLIP-drug conjugates to ameliorate lung injury. Taken together, the results 2 provide new insights into the molecular pathology of influenza pneumonia, and offer opportunities to improve the management of influenza-induced lung disease 3 List of figures Figure 1.1 Structure of an influenza virus. ..................................................... 10 Figure 1.2 The life cycle of influenza virus. .................................................... 11 Figure 1.3 Inflammation-induced ROS / RNS damages cellular molecules. . 25 Figure 1.4 Structures of modified nucleobases upon exposure to ROS / RNS................................................................................................................... 28 Figure 1.5 Classical DNA double strand break repair pathways. .................. 36 Figure 3.1 PR8 infection of MDCK cells induced H2AX foci formation. ...... 73 Figure 3.2 Disease progression during sublethal PR8 infection................... 78 Figure 3.3 Characterization of inflammation kinetics. ................................... 83 Figure 3.4 Real-time PCR analysis of IFN- in lung homogenate. ................. 84 Figure 3.5 Oxidative stress is elevated especially when viral load is suppressed. ..................................................................................................... 85 Figure 3.6 Reduction in superoxide dismutase (SOD) activity on 7 dpi. ...... 86 Figure 3.7 Western analysis of influenza antigen and H2AX. ...................... 87 Figure 3.8 DSB markers, H2AX and 53BP1 foci, were induced in bronchiolar epithelium cells after active infection. ....................................... 89 Figure 3.9 H2AX foci were induced in alveolar epithelial type II cells in the recovery phase of influenza infection. ........................................................... 91 Figure 3.10 Increased H2AX foci formation in both infected and uninfected cells. ................................................................................................................. 92 Figure 3.11 Pan-nuclear H2AX is found in the same regions as cleaved caspase 3 positive cells in sequential sections. ........................................... 93 Figure 3.12 H2AX foci formation increases in immune cells after infection. .......................................................................................................... 95 Figure 3.13 PARP mediated PAR polymerization increases after infection . 97 Figure 3.14 PARP-1 is cleaved following influenza infection ........................ 99 Figure 3.15 DNA damage is initiated before, and proceeds after whole lung apoptosis. ...................................................................................................... 101 4 Figure 3.16 H2AX phosphorylation increases in Ki-67+ cells after infection. ........................................................................................................ 104 Figure 3.17 NHEJ and HR-related genes are induced during tissue regeneration................................................................................................... 107 Figure 3.18 HR-related genes are induced during the later stage of infection. ........................................................................................................ 108 Figure 3.19 DNA-PKcs activity is dispensable for DNA repair in bronchial epithelium on 9 and 13 dpi............................................................................ 111 Figure 3.20 PR8 infected mice injected with 2 doses of NU7441 at 9 dpi. .. 112 Figure 4.1 RaDR-GFP HR substrate recombines to encode for full length EGFP sequences. .......................................................................................... 127 Figure 4.2 PCR detection of HR that led to full length EGFP reconstitution in RaDR-GFP pancreatic cells .......................................................................... 135 Figure 4.3 Single strand annealing does not reconstitute RaDR-GFP........ 136 Figure 4.4 RaDR-GFP substrate yields different recombination products following gene conversion, sister chromatid exchange, and replication fork repair. ............................................................................................................. 138 Figure 4.5 Nested PCR analysis of single cells could identify cells that had undergone crossover associated-HR or gene conversion without crossover. ...................................................................................................... 140 Figure 4.6 EGFP positive cells arise in the lung epithelial cells indicating HR at Rosa26 locus ............................................................................................. 142 Figure 5.1 pHLIP peptide sequences. ........................................................... 147 Figure 5.2 pHLIP exists in 3 states and facilitates bidirection transport of cargo molecules. ........................................................................................... 150 Figure 5.3 pHLIP preferentially binds cell membrane at acidic pH in vitro. 151 Figure 5.4 pHLIP targets infected but not uninfected mice lungs............... 155 Figure 5.5 pHLIP targets acidic tissue. ......................................................... 157 Figure 5.6 Bioimaging of pHLIP-induced fluorescence on mice ventral surface to estimate peptide clearance. ........................................................ 159 Figure 5.7 pHLIP is retained in infected lungs 8 days post peptide injection. ........................................................................................................ 160 Figure 5.8 K-pHLIP accumulates in infected lungs via pH-independent mechanism..................................................................................................... 161 5 Figure 5.9 pH responsiveness of pHLIP variants, Leu26Gly, K-pHLIP and D3K-pHLIP...................................................................................................... 164 Figure 5.10 Evidence of pH-dependent targeting of pHLIP and Leu26Gly. 167 Figure 5.11 pHLIP targets heavily infiltrated regions of lungs. ................... 170 Figure 5.12 pHLIP accumulates in lungs during the later phase of infection. ........................................................................................................ 172 Figure 5.13 pHLIP accumulates in damaged lung tissue. ........................... 174 Figure 5.14 pHLIP localizes at lung tissue depleted of Prdx6. .................... 176 Figure 5.15 Quantification of pHLIP-induced pixels in Prdx6 positive and negative regions. ........................................................................................... 177 6 List of tables Table 1.1 Licensed influenza anti-viral agents ............................................... 16 Table 2.1 Commercial sources of media and reagents ................................. 42 Table 2.2 Formulations for solutions and buffers ......................................... 43 Table 2.3 Sources and clones of antibodies used ......................................... 44 Table 2.4 Panel 1 antibodies for myeloid cells .............................................. 45 Table 2.5 Panel 2 antibodies for lymphoid cells ............................................ 45 Table 2.6 Sequences and molecular weights of pHLIP and peptide variants ............................................................................................................ 46 Table 2.7 Antigen retrieval methods ............................................................... 51 Table 2.8 Mean number of Ki-67+ and H2AX+ cells counted in 15 40x magnified images ............................................................................................ 54 Table 2.9 PCR primers to specifically amplify full length EGFP, Δ3egfp, and Δ5egfp. ............................................................................................................. 59 Table 2.10 External PCR primers designed to anneal upstream and downstream of the EGFP coding sequence. ................................................. 60 Table 2.11 Thermal cycler conditions and product sizes. ............................. 60 Table 3.1 Summary of DNA damage in infectious diseases ......................... 68 Table 3.2 Interplay between viral infection and DDR pathway that is independent of DNA damage .......................................................................... 69 Table 3.3 Pathology analysis of mice lung sections following PR8 infection. .......................................................................................................... 79 Table 5.1 Anti-inflammatory agents and their reported efficacy on influenza infection ......................................................................................................... 146 Table 5.2 Alterations in sequences and Molecular weight of pHLIP variants .......................................................................................................... 161 Table 5.3 Predicted aggregation score and reported pH-sensitivity .......... 163 7 Chapter 1 Background 1.1 Influenza virus Influenza viruses (commonly known as flu) are airborne viruses from the orthomyxoviridae family. There are three genera (or antigenic type) of influenza viruses in this family -Influenza A, Influenza B and Influenza C – among which only influenza A and influenza B are known to cause severe respiratory illnesses and epidemic outbreaks in the world. In contrast, Influenza C give rise to mild respiratory diseases in human and is not thought to contribute to annual epidemics (Zambon 1999). Influenza viruses have been commonly found to be roughly spherical in shape, and are made up of lipid-bilayer envelopes derived from host plasma membrane. Projecting out of the surface of each influenza virion, are approximately 500 spike – liked glycoproteins, haemagglutinin (HA) and neuraminidase (NA), that make up ~ 80% and ~ 17% of all viral envelope proteins respectively. Around 16 to 20 molecules of another minor viral antigen, matrix 2 protein (M2), can also be found on the envelope of each virion. Right underneath the viral envelope, there is a spread - out distribution of matrix 1 proteins (M1), which are responsible for binding viral ribonucleoprotein (vRNP) complexes located in the core of each virus particle (Figure 1.1) (Samji 2009; Ruigrok et al. 1984). There are eight rod-shaped vRNP complexes in a virion, and each vRNP complex is composed of a viral RNA associated with multiple nucleoproteins (NP) and a trimeric viral polymerase complex. The 5’ and 3’ termini of viral RNA are partially complementary, allowing vRNP to form a helical corkscrew structure. Influenza viruses possess eight segmented, single-stranded, negative sense RNA genomes 8 approximately 13 thousand bases in size. Each RNA segment forms a single vRNP with NP monomers and viral polymerase complex in a mature virion. These eight segments are historically known to encode for 10 viral proteins, namely three polymerases (PA, PB1, PB2), NP, M1 and M2, HA, NA and non-structural proteins (NS1 and NS2) (Figure 1.1). More recently, influenza has also been shown to express up to 16 viral proteins, which include PB1-F2 (encoded by the +1 alternate open reading frame within PB1 gene), a M2-related protein M42, N40 (encoded by PB1), PA-X, PA-N155 and PA-N182 (encoded by PA) (Shi et al. 2014; Chakrabarti and Pasricha 2013). Influenza viruses are highly diverse pathogens, with different viruses possessing varied transmissibility, pathogenicity and infectivity in hosts. Many subtypes of Influenza A viruses exist, and they are classified based on their expression of different HA and NA antigens. Up to early 2015, a total of 18 HA and 11 NA antigens have been identified among influenza A viruses isolated from humans, birds and bats (Freidl et al. 2015). Based on the expression of different HA and NA antigens, Influenza A viruses can be broadly categorized into subtypes such as H1N1, H3N2 and H5N1, which can then be further subdivided into strains. On the contrary, there is only one subtype of Influenza B, though it can also be further subdivided into lineages and strains, much like Influenza A (CDC1). Based on a commonly accepted naming convention published in the Bulletin of the World Health Organization (WHO) during 1980, individual Influenza strains are identified by the antigenic type, the host of origin, the geographical origin where the strain was isolated, the strain number, the year of isolation, followed by the HA and NA 1 Transmission of Influenza Viruses from Animals to People. (2014, August 19). Centers for Disease Control and Prevention. Retrieved June 01, 2015, from http://www.cdc.gov/flu/about/viruses/transmission.htm. 9 antigen description in parentheses for Influenza A viruses [e.g. Influenza A/Puerto Rico/8/37 (H1N1) and Influenza B/Yamagata/16/88] (WHO 1980). Figure 1.1 Structure of an influenza virus. Influenza is an enveloped virus composed of two major surface glycoproteins, haemagglutinin (HA) and neuraminidase (NA), and a minor component M2. The genome of influenza consists of 8 RNA segments, which are folded into ribonucleoprotein complexes (RNP) and encode for nucleoprotein (NP), three polymerase proteins (PA, PB1, and PB2), matrix proteins (M1 and M2), nonstructural proteins (NS1 and NS2) and 2 glycoproteins (HA and NA). (Image is taken from Nelson MI, Holmes EC. The evolution of epidemic influenza.Nat Rev Genet 2007 Mar;8(3):196-205) 1.1.1 Life cycle of Influenza virus The life cycle of an influenza virus occurs in several steps, which will be divided into five main stages here to facilitate description: (1) viral binding and entry into host cells, (2) uncoating and vRNPs translocation into the nucleus, (3) transcription and replication of viral RNA, (4) synthesis of new vRNPs and finally, (5) assembly and budding (Figure 1.2). 10 Figure 1.2 The life cycle of influenza virus. When influenza virus infects a host, viral envelope protein HA first binds to sialylated host receptors that facilitate the entry of the virus into host cells via receptor – mediated endocytosis. Endosomal acidity then triggers the fusion of viral and endosomal membranes, resulting in uncoating of viral membrane and release of vRNP complexes into the cytosol. Subsequently, vRNP complexes translocate into host cell nucleus, where viral RNA segments are transcribed and replicated. In the cytoplasm, host translational machinery is exploited to synthesize viral proteins including NP, PA, PB1 and PB2, which are then translocated into the nucleus to assemble into novel vRNPs along with newly synthesized viral RNA. These newly generated vRNPs are then transported out of the nucleus with the help of M1 and NS2. Finally, viral components are assembled at the cell membrane, where progeny virus buds out at the apical surface of host plasma membrane. (Image is taken from Shi Y, Wu Y, Zhang W et al. Enabling the 'host jump': structural determinants of receptor-binding specificity in influenza A viruses. Nat Rev Microbiol. 2014 Dec;12(12):822-31.) (1) Influenza infection is initiated with the binding of HA to sialic acid-linked glycoproteins (sialic acid receptors) expressed on host cell membrane. This 11 step is thought to be an important determinant of host infectivity since avian influenza viruses generally binds to α 2, 3 - linked (avian-type) sialic acid receptors that are dominantly found in the avian respiratory and gastrointestinal tracts, while human influenza viruses preferentially bind to α 2, 6 - linked (human-type) sialic acid receptors that are prevalent in the human airways. Other animals, such as the pigs, Japanese quail and mice possess both α 2, 3 - linked and α 2, 6 - linked sialic acid receptors, and hence, can be infected by both avian and human influenza viruses [Reviewed in (Kimble, Nieto, and Perez 2010; Ning et al. 2009)]. Following binding, Influenza virus enters the host cell inside an endosome via receptor-mediated endocytosis. (2) Acidity in the endosome (~pH 5 - 6) leads to a change in the conformation of HA trimer, allowing HA2 (a subunit of HA) to facilitate fusion of the viral and endosomal membranes, leading to the formation of a pore which gives passage for vRNP complexes to enter the cytosol. M2, a proton ion channel, further pumps H+ into the viral core, decreasing the pH of the viral core and disrupting the interactions between vRNP complexes and M1, so that vRNP particles can now freely move into the cytoplasm. Nuclear localization signals found in vRNP proteins (NP, PA, PB1 and PB2) then mediates the entry of vRNP complexes into host nucleus using host cell’s nuclear import machinery. (3) In the nucleus, negative sense viral RNA are then transcribed into positive sense messenger RNAs (mRNAs) and complementary RNAs (cRNAs) using the trimeric viral RNA- dependent RNA polymerase (RdRp) composed of PA, PB1 and PB2. PB2, containing endonuclease activity, binds and cleaves 5’ methylated caps of host mRNAs via a “cap- snatching” mechanism to prime viral transcription. As such, viral mRNAs will possess 5’ methylated caps even 12 though 5’ caps are not encoded in the viral genome. Newly synthesized mRNAs are subsequently transported into the cell cytosol to be translated into viral proteins by exploiting the host translational machinery. cRNAs remain in the nucleus as templates for synthesis of multiple copies of negative sense viral RNA. (4) Newly synthesized NP and viral polymerases are transported back into the nucleus, where they can assemble with negative sense viral RNA to form vRNPs. These negative sense vRNPs are subsequently exported out of the nucleus through nuclear pores via exportin-1 (also known as CRM1) – mediated nuclear export, potentially via their interaction with M1 and NS2, which will then bind to exportin-1 (Samji 2009; Elton et al. 2001) . (5) In order for budding to occur, influenza components must first assemble at the plasma membrane. Following protein synthesis in the endoplasmic reticulum, envelope antigens HA, NA and matrix proteins are transported by the transGolgi network to host plasma membrane. Newly synthesized vRNPs also bind to M1 that are assembled at the inner surface of host membrane bilayer, preventing the re-entry of vRNPs into the nucleus. If influenza replication occurs in the respiratory tract, viral components assemble at the apical side of polarized epithelial cells where progeny viruses will exclusively emerge from, so that they can re-infect other epithelial cells in the airway, instead of entering the circulatory system. After budding, progeny viruses are still bound to host cell membrane via HA and sialic-receptor interaction. NA (and other host proteases) then cleaves sialic acid residue from surface glycoproteins and glycolipids to release mature virions into the surrounding, completing the life cycle of an influenza virus (Samji 2009; Bottcher et al. 2006). 13 1.1.2 Antigenic drift and antigenic shift Influenza is a highly successful virus. Based on historical evidences, influenza viruses have circulated in the human population since more than half a millennium ago (Taubenberger and Morens 2010), and yet, despite the advent in vaccines and anti- viral therapies, it is still impossible to eradicate influenza virus. Influenza A is commonly known to be the most virulent genus of influenza viruses, and has been found to infect many animals such as birds, human, swine, horses, whales, seals and bats, among which wild birds are thought to be the primary natural reservoir of influenza A virus (Spackman 2009). The success of influenza A virus can be partly attributed to its error-prone RdRp, which makes incorporation errors at an estimated rate of 7.2×10−5 bp−1 per replication cycle (approximately 1 mistake per genome per cycle) (Drake 1993). High error rate in genome replication allows influenza A viruses to gradually mutate and develop modified antigens that escape recognition by pre-existing host immunity. This process, called antigenic drift, is usually the cause of recurrent epidemics within communities which do not have specific immunity against the new virus. The ability of influenza A viruses to mutate also allows them to develop resistance towards currently available anti-viral agents, thereby creating challenges for treating influenza - infected patients. There are currently only five licensed influenza anti-viral agents in the United States of America (CDC2), out of which only three NA inhibitors, Tamiflu® (Oseltamivir phosphate), Relenza® (Zanamivir) and Rapivab® (Peramivir), are recommended by U.S. Food and Drug 2 Influenza Antiviral Medications: Summary for Clinicians. (2015, February 25). Centers for Disease Control and Prevention.Retrieved on 30 May 2015, from http://www.cdc.gov/flu/professionals/antivirals/summary-clinicians.htm. 14 Administration (FDA). The other two M2 blockers, Symmetrel® and Flumadine®, have been in use for therapeutic purposes for a longer duration of time, but are no longer recommended as prescriptions for influenza therapy and prophylaxis. This is because amantadine and rimantadine, the compound names for Symmetrel® and Flumadine® respectively, are found to be ineffective during recent outbreaks with a widespread emergence of resistant H3N2 and 2009 pandemic H1N1 viruses, such that more than 99 % of all strains tested are resistant to both compounds. In contrast, Tamiflu®, Relenza® and Rapivab® (recently approved agent) are still largely effective for the control of seasonal H3N2 and 2009 pandemic H1N1 viruses (more than 98% of viruses tested are susceptible). However, H1N1 viruses resistant towards Tamiflu® and Rapivab® due to single amino acid mutations (H274Y or N294S alterations) in the NA protein have also began to emerge (Hurt et al. 2009; Pizzorno et al. 2011) (Table 1.1). Given the general effectiveness of Tamiflu® and Relenza® towards recent circulating strains, the Health Science Authority of Singapore recommends prescribing these two drugs for treating influenza infected patients. 15 Table 1.1 Licensed influenza anti-viral agents (Information taken from CDC3 and Health Science Authority4) Antivirals Action Routes of administration Resistance and effectiveness Tamiflu® (Oseltamivir phosphate) NA inhibitor Oral - - Relenza® (Zanamivir) NA inhibitor Inhaled - Rapivab® (Peramivir) NA inhibitor Intravenous - Symmetrel® (Amantadine) M2 blocker Oral - Flumadine® (Rimantadine) M2 blocker Oral - 3 Less than 2% of tested 2009 H1N1 viruses were resistant to H1N1 Not effective towards Influenza A strains with H274Y or N294S alterations Effective with Influenza A H3N2 and Influenza B Approved for adults and children by Health Science Authority 100% of 2009 H1N1 viruses, H3N2 and influenza B were susceptible to Zanamivir Recently approved by FDA in December 2014 Not effective towards Influenza A strains with H274Y or N294S alterations Only approved for adult treatment by Health Science Authority Effective only for influenza A, but not influenza B High resistance (> 99%) in recent circulating H3N2 and 2009 pandemic H1N1 Effective only for influenza A, but not influenza B High resistance (> 99%) in recent circulating H3N2 and 2009 pandemic H1N1 Influenza Antiviral Drug Resistance. (2015, January 8). Centers for Disease Control and Prevention.Retrieved on 07 June 2015, from http://www.cdc.gov/flu/about/qa/antiviralresistance.htm. 4 Frequently Asked Questions. (2014, April 22). Singapore Health Sciences Authority (HSA).Retrieved on 07 June 2015 from http://www.hsa.gov.sg/pub/faq/faq/faqcategory/antiviral-drugs-what-are-they-.aspx . 16 Besides the occurrence of random mutations in influenza genome, Influenza virus possesses eight segregated RNA segments in its genome that enable genetic reassortment to take place. This happens when two or more influenza subtypes co-infect a single host (e.g. swine) that expresses both α 2, 3 - linked and α 2, 6 linked sialic acid receptors, and their RNA segments mix to form new combinations for novel viruses. Reassortment introduces drastic modification to the virus genome, and hence the phenotype, causing a process called antigenic shift. By replacing the surface NA and HA antigens of influenza A viruses from an animal origin to those of a human origin, novel strains that can infect humans are created, and may lead to large scale pandemic outbreaks (Treanor 2004). It is difficult to accurately predict which novel subtypes will emerge. Hence, while annual vaccination of updated influenza vaccines is the best strategy of protecting oneself from influenza infection, these influenza vaccines may not confer enough protection against novel strains (Plans-Rubio 2012). Furthermore, it takes at least five to six months lead time to generate a customized vaccine after identifying the pandemic strain, thereby imposing a huge problem for health care systems when outbreaks occur (Haaheim, Madhun, and Cox 2009). Taken together, genetic variations that give rise to novel influenza strains are the greatest hurdle for effective vaccines and antiviral drugs. With the continual circulation of influenza in the animal and human populations, this virus is expected to be a public health concern for a long time to come. 1.1.3 Global burden imposed by influenza outbreaks One can be easily infected with influenza viruses upon short range contact (within two meters) with infected individuals through respiratory droplets from a day before 17 symptoms appear, to a week or more after they become sick (CDC5). If caution is not exercised, it is also possible to spread flu virus through contaminated objects. Influenza usually manifests in the upper airway with mild cold-like symptoms that can subside with sufficient rest and fluids. However, it can sometimes cause severe to life-threatening pneumonia and other complications that require intensive medical attention. Severe influenza infection is especially prevalent among children and elderlies, pregnant women and people with pre-existing medical conditions (Jain et al. 2009; Ljungman et al. 1993). There is also a particularly high risk of influenza - infected patients succumbing to bacteria co-infection (Chertow and Memoli 2013). Based on an estimation made by the WHO, three to five millions of people worldwide suffer from severe debilitating influenza infections every year, while approximately 250-500 thousand people succumb and die from seasonal influenza and influenzaassociated complications6. A more worrisome scenario arises with the emergence of zoonotic influenza strains such as the H5N1 and H7N9 bird flu which have exceedingly high mortality rates of approximately 60% and 18.7% (Mei et al. 2013). It has been estimated that more than 60 million (up to 81 million) people may die today, should an influenza pandemic of the same scale as the 1918 Spanish flu outbreak occur (Murray et al. 2006). Together with escalating costs in treatment fees, and monetary loss from reduced work productivity and absenteeism, influenza outbreaks continue to bring about drastic socio-economic impacts on individuals and society (Szucs 1999). 5 How Flu Spreads. (2013, September 12). Centers for Disease Control and Prevention.Retrieved July 22, 2014, from http://www.cdc.gov/flu/about/disease/spread.htm. 6 Influenza (Seasonal) Fact Sheet (2014, March). World Health Organization.Retrieved July 22, 2014, from http://www.who.int/mediacentre/factsheets/fs211/en/. 18 1.2 Influenza infection induces acute lung injury The onset of severe illnesses due to influenza infection is often due to infection spreading from the upper respiratory tract to the lower respiratory tract. A proportion of patients with lower respiratory tract infection develop progressive pneumonia with gross lung damage and impaired oxygen intake. This is medically known as acute lung injury (ALI) and acute respiratory distress syndrome (ARDS), characterized by acute infiltration of immune cells into lungs, and a reduction of oxygen level in the blood stream (hypoxemia) (Johnson and Matthay 2010). Influenza patients with ALI / ARDS are typically presented with diffused alveolar damage, bronchiolitis and organizing pneumonia. In many cases, hemorrhage, edema, hyaline membrane formation and fibrosis can also be observed (Fujita et al. 2014; Nakajima et al. 2012). As a result, lung functions are greatly compromised. Alveolar epithelial and endothelial cell death during the course of infection is a leading cause of lung failure. The presence of infiltrated cells and fluid (edema) in the lung further diminishes air volume in the lungs, thereby enhancing hypoxemia. Depletion of blood oxygen can cause multiple organs to fail since the entire body relies on oxygen to survive. For instance, liver, which is positioned away from the site of infection (lungs), shows signs of hepatocellular injury during severe influenza insult, potentially due to hypoxia, and exposure to pro-inflammatory mediators in the systemic circulation (Papic et al. 2012; Han et al. 2014). Clinical statistics have shown that influenza patients with ALI / ARDS, such as those infected with pandemic H1N1 (2009), have a high reported mortality rate of 15% in adults and children (Quispe-Laime et al. 2010; Ali et al. 2013), despite availability of medical attention. Hence, it is important to develop novel targets and strategies 19 that can be exploited to overcome the limitations of current influenza therapy, in order to reduce fatalities during severe influenza infections. 1.2.1 Influenza-induced cytopathy and lung injury Influenza virus is a causative agent for significant cell death. In particular, programmed cell death (apoptosis) is considered to be a hallmark of influenza infection and is thought to contribute to lung injury during influenza pneumonia (Hashimoto et al. 2007; Mori et al. 1995; Short et al. 2014). In cells, multiple components of influenza such as NP, M2, PB1-F2 and NS1 have been demonstrated to cause predominantly mitochondria-dependent apoptosis as a viral strategy to increase virus replication and release (Tripathi et al. 2013; Zamarin et al. 2005; Schultz-Cherry et al. 2001). Consistent with this finding, inhibition of pro-apoptotic factors such as BAD and Bax dramatically reduces viral titer in cell cultures (Tran et al. 2013; McLean et al. 2009), highlighting that the mechanisms involved in host cell apoptosis is exploited by influenza virus to propagate. Viralinduced cell death can be said to be a direct cause of pathogenesis during influenza infection. Hence, anti-viral agents that can block viral entry or replication, such as oseltamivir, remain a key strategy in containing the spread and detrimental effects of influenza. 1.2.2 Inflammatory responses and lung injury It is now clear that attributing influenza-induced lung injury solely to viral activity does not adequately describe the disease process. In response to infection, influenza infection elicits a classical innate to adaptive inflammatory response (Buchweitz, Harkema, and Kaminski 2007; Pommerenke et al. 2012) to increase host resistance towards influenza (Mordstein et al. 2010), to suppress viral 20 replication (Tumpey et al. 2005) and to remove viral-infected cells via host-cell apoptosis (Topham, Tripp, and Doherty 1997; Brincks, Katewa, et al. 2008; Brincks, Kucaba, et al. 2008; Ishikawa et al. 2005). While the immune systems have evolved to protect hosts from pathogens, paradoxically, there is evidence that host responses can sometimes contribute to increased pulmonary injury and even quicken death in animal models of severe influenza infection. Existing data have shown that excessive inflammatory responses such as cytokine storm, high neutrophil influx, uncontrolled T cell cytotoxicity and unmodulated inflammatory signaling have been implicated with more severe disease outcomes (Brandes et al. 2013; de Jong et al. 2006; Sakthivel et al. 2014). Although there is a conflicting school of thought which proposed that heightened inflammation is merely a consequence of poorly contained viral load, and that lung injury is mainly attributable to high virus titers (Boon et al. 2011), there are numerous animal data which support the hypothesis that excessive inflammation drives pathogenesis during infection. Specifically, empirical studies demonstrated that attenuation of pro-inflammatory cells and molecules improves lung pathology and survival following influenza infection (Berri et al. 2013; Snelgrove et al. 2008; Walsh et al. 2011; Lauder et al. 2011). Taken together, these data underscore the detrimental effects of excessive inflammatory responses on tissue integrity during infection. 1.2.3 Reactive oxygen and nitrogen species during influenza infection Reactive oxygen species (ROS) and reactive nitrogen species (RNS) are an integral component of inflammation. Activated neutrophils and macrophages rapidly consume oxygen and glucose to undergo respiratory burst through the activation of NADPH oxidase, and involvement of other enzymes such as 21 myeloperoxidase (abundant in neutrophils) and inducible nitric oxide synthase (iNOS). Respiratory burst results in the release of several ROS / RNS including superoxide (O2−•), hydrogen peroxide (H2O2) and nitric oxide (NO). O2−• can be further dismutated into H2O2, which can then be converted to highly reactive hypochlorous acid (HOCl) by myeloperoxidase (Robinson 2008; Lonkar and Dedon 2011; Buffinton et al. 1992). In combination, these ROS / RNS exert strong microbicidal properties, and some of them (e.g. NO and O2−•) can also serve as second messengers during signal transduction of normal physiological processes (Forman and Torres 2001). It was previously reported that leukocytes (Gr1+ and CD11b+ myeloid cells such as neutrophils and monocytes) extracted from influenza - infected mice lungs contain higher levels of ROS and nitrotyrosine (nitrosylated tyrosine) as compared to those taken from uninfected mice, suggesting an increase in ROS / RNS production by the myeloid cells (Lee et al. 2013). Under normal physiological condition, ROS / RNS can be removed rapidly by intracellular antioxidants and radical scavengers to dissipate their damaging effects. However, evidence has shown that influenza infection causes downregulation of pulmonary antioxidants such as glutathione, superoxide dismutase (SOD), catalase and vitamin E, which can drive further redox imbalance (Hennet, Peterhans, and Stocker 1992; Kumar et al. 2005; He et al. 2013). In addition, other components of the inflammatory process such as tumor necrosis factor-α (TNF-α) and granzymes / perforin are also involved in augmenting intracellular oxidative stress in cells (Wheelhouse et al. 2003; Suematsu et al. 2003; Martinvalet et al. 2008), and may contribute to elevated oxidative stress in lung tissue. Indeed, direct measurement by electron spin resonance spectroscopy 22 revealed that higher levels of ROS / RNS and oxidized biomolecules (e.g. malondialdehyde) are present in severely infected mouse lungs (H5N1 and H2N2 models) as opposed to uninfected mouse lungs (He et al. 2013; Akaike et al. 1996). Although ROS / RNS are crucial molecules involved in regulatory cell signaling and anti-microbial activities, an excess of ROS / RNS can cause tissue injury (Chabot et al. 1998). In animals infected with influenza, it was found that the administration of enzymatic (e.g. SOD) or small molecule antioxidants suppresses lung injury (Oda et al. 1989; He et al. 2013; Snelgrove et al. 2006; Akaike et al. 1996), suggesting that excessive ROS / RNS leads to more severe outcomes. Likewise, deficiency or inhibition of xanthine oxidase (XO, a O2−• generating enzyme), Nox2 (catalytic subunit of NADPH oxidase found on phagocytes), or iNOS (induced on activated phagocytes), have been shown to dramatically reduce immune cell infiltration, the levels of pro-inflammatory mediators, cell death and mortality rates (Vlahos et al. 2011; Akaike et al. 1990; Karupiah et al. 1998). Surprisingly, treatment with antioxidants has little or no effect on virus titers despite dampening inflammatory responses, and mice with iNOS or Nox2 gene knockout even have reduced viral loads as compared to wildtype mice (Vlahos et al. 2011; Karupiah et al. 1998; Snelgrove et al. 2006). Interestingly, while deficiency in Nox2 confers benefits to hosts during influenza infection, some evidence suggests that deficiency in the Nox1 isoform of NADPH oxidase leads to a slight elevation of lung inflammation and increased weight loss. The authors found small increases in neutrophil infiltration during the acute phase of inflammation and heightened levels of pro-inflammatory cytokines in Nox1 knockout mice infected with influenza virus. Notably, the deficiency of Nox1 during influenza infection increases the ability of immune cells to generate ROS upon stimulation (Selemidis et al. 2013). Nox1 and 23 Nox2 (also known as gp91phox) are homologs of the catalytic subunits found in non-phagocytic NOX (NOX1) and phagocytic NOX (NOX2) respectively. While Nox1 and Nox2 are extremely similar in structure, Nox1 are found to be expressed by lung epithelial and endothelial cells, whereas Nox2 are specifically expressed by phagocytes (Selemidis et al. 2013). Although it is not clear why Nox2 deficiency protects mice against influenza infection while Nox1 deficiency increases injury, one possibility could be that small amounts of ROS generation by NOX1 in lung epithelial and endothelial cells is necessary for modulating the expression of proinflammatory cytokines and Nrf2 (Selemidis et al. 2013), such that when NOX1 is inhibited, phagocytes are primed to produce large amounts of ROS by NOX2mediated respiratory burst. Taken together, extensive experimental evidence suggests that while some level of ROS / RNS generated by lung cells is essential for maintaining normal lung physiology, and may be involved in controlling inflammation, excessive ROS / RNS induced by the inflammatory systems can exacerbate lung injury during influenza infections. It has been thought that oxidative / nitrosative stress can cause tissue injury by modifying biomolecules, which in turn alter cell signaling involved in regulating endothelium barrier, cell death and tissue reparation (Mittal et al. 2014; Sunil et al. 2012). However, the mechanisms underlying ROS / RNS-induced pathogenicity during influenza infections are not yet fully understood. Under normal physiological conditions, the body produces low concentrations (nanomolar) of endogenous ROS / RNS (e.g. O2−• and NO) that are short lived and have limited reactivity. However, when high concentration of ROS / RNS are generated during inflammation, they may not be removed quickly, and can be converted to highly reactive molecules including hydroxyl radicals (OH•), lipid peroxides (ROOH) and 24 peroxynitrite (ONOO-) (Mittal et al. 2014) (Figure 1.3). These toxic molecules can react with almost any biomolecule including lipids, proteins and nucleic acids in the body and are capable of altering cell physiology and causing cell damage (Beckman and Koppenol 1996; Lonkar and Dedon 2011; Dedon and Tannenbaum 2004). For instance, ROS / RNS can induce post-translational modifications such as S-nitrosylation, tyrosine nitration, carbonylation and disulfide linkage to proteins. One example of such modifications is the oxidative modifications to cofilin, which then promotes mitochondria swelling and cytochrome c release that eventually leads to apoptosis (Klamt et al. 2009). Similarly, irreparable oxidative damage to other essential biomolecules, such as lipids and nuclei acids, can result in cell injury (Tribble, Aw, and Jones 1987; Lonkar and Dedon 2011). Given that oxidative stress is elevated during influenza infection, a potential mechanism of lung injury is oxidative damage to cellular molecules during the inflammatory process. Understanding the implications of influenza infection on the integrity of cellular molecules may shed light on the most fundamental mechanisms that contribute to tissue injury during influenza infection. Figure 1.3 Inflammation-induced ROS / RNS damages cellular molecules. 25 Activated neutrophils and macrophages generate ROS / RNS that further react to form toxic compounds that modify biomolecules such as proteins and DNA. (Image taken from Dedon PC, Tannenbaum SR. Reactive nitrogen species in the chemical biology of inflammation. Arch Biochem Biophys. 2004 Mar 1;423(1):12-22.) 1.3 ROS and RNS can damage genomic DNA Genomic DNA is an important target of intracellular ROS / RNS. Every component of the DNA, including its nucleobases, deoxyribonucleosides and its ribosephosphate backbone can be damaged by reacting with various ROS / RNS. Under normal physiological conditions, energy metabolism and other biochemical reactions contribute to continuous production of ROS / RNS that leads to a low steady-state level of DNA damage. Accumulation of DNA damage over time has been proposed to play roles in promoting slow biological processes such as neurodegenerative diseases and aging (De Bont and van Larebeke 2004; Maynard et al. 2009). With the onset of inflammation, excessive formation of ROS / RNS including H2O2, HOCl, singlet oxygen, ONOO- and OH• will overwhelm antioxidant systems and exacerbate DNA damage. OH• radical is considered one of the most potent ROS that leads to DNA damage by oxidizing bases and attacking DNA backbones to generate DNA strand breaks. Due to its reactivity, OH• radical does not diffuse beyond one or two molecular diameter without reacting with a neighboring molecule. Hence, OH• needs to be near to DNA in order to induce any modifications to the genomic molecules(Pryor 1986). It is thought that H2O2, a membrane permeable and more diffusible oxidant, can diffuse into nucleus to be close enough to DNA molecules. H2O2, commonly generated under inflammatory conditions, then give rise to DNA-damaging OH• via Fenton reaction, although H2O2 itself does not appear to directly cause DNA damage. The reduction of H2O2 to OH• takes place in the presence of transition 26 metals such as copper Cu(II) and iron Fe (III) found as co-factors in nuclear proteins (reviewed in (Cadet et al. 1999; Wiseman and Halliwell 1996; Dizdaroglu 1992)), as shown in the equation for Fenton reaction below: Fe2+ + H2O2 + H+  Fe3+ + H2O + OH•. To date, more than 100 types of oxidative nucleobases and deoxyribose modifications have been identified after DNA treatment with H2O2 and other ROS / RNS (Dizdaroglu 1992; Croteau and Bohr 1997). Modified nucleobases are highly diversified (some examples shown in Figure 1.4), among which 8-hydroxydeoxyguanine (8-OH-dG) is one of the most prevalent and certainly, the most studied form of oxidative base lesion (Dizdaroglu 1992). In addition, other modifications exist to add on to the spectrum of DNA damage when DNA reacts with inflammation-induced reactive species and reactive intermediates. For instance, DNA’s interaction with HOCl can give rise to halogenated nucleobases (e.g. 8-chloro(2’-deoxy)guanosine, 5-chloro(2’-deoxy)cytidine, and 8-chloro(2’deoxy)adenosine) (Masuda et al. 2001), while reaction with dinitrogen trioxide, a RNS, has also been shown to cause base deamination to form 2’-deoxyxanthosine, 2’-deoxyoxanosine, 2’-deoxyinosine and 2’-deoxyuridine (Dong et al. 2006). In addition, intermediate products of oxidation, such as lipid peroxides (e.g. 4hyrdoxynonenal and malondialdehyde) can also induce damage in DNA to form strongly mutagenic etheno-DNA adducts (Linhart, Bartsch, and Seitz 2014). Taken together, numerous forms of DNA modification can occur upon exposure to ROS / RNS, including abasic sites, bulky adduct, oxidized bases, deaminated bases, and DNA- DNA and DNA- protein cross-links (Lonkar and Dedon 2011). In addition, attack of ROS / RNS on DNA backbone can also lead to DNA single strand breaks (SSBs), which can further give rise to more toxic DNA double strand breaks (DSBs) 27 when two SSBs occur at close proximity to each other in complementary strands, or when a replication fork collapse at a SSB. These DNA lesions are responsible for base mispairing, loss of genetic information, and can inhibit DNA replication and transcription (Jackson and Bartek 2009). Figure 1.4 Structures of modified nucleobases upon exposure to ROS / RNS. Some examples of base moieties in ROS / RNS-damaged deoxynucleotides are shown in the figure. For instance, 8-hydroxyguanine and 2, 6-diamino-4-hydroxy-5formamidopyrimidine (Fapy-Gua) are two of the most abundant DNA base modifications found in biological systems. Besides oxidized bases, reaction with reactive nitrogen species can produce deaminated bases such as xanthine from guanine, and hypoxanthine from adenine. (Image is taken from Kamiya H. Mutagenic potentials of damaged nucleic acids produced by reactive oxygen/nitrogen species: approaches using synthetic oligonucleotides and nucleotides: survey and summary. Nucleic Acids Res. 2003 Jan 15;31(2):517-31.) 28 1.3.1 Cellular responses towards DNA damage When DNA lesions are unrepaired or repaired improperly, toxicity or wide-scale genome rearrangements can occur, affecting cell or organism survival (Jackson and Bartek 2009; Helleday et al. 2007). Mispairing of damaged nucleotides causes a change in genetic information that promotes the formation of mutations. For instance, when DNA replicates, 8-hydroxy-deoxyguanosine and Fapy-Gua have been shown to mispair with adenine leading to a G: C to A: T transversion, a common somatic mutation found in human cancers (Boiteux and Radicella 1999; Gehrke et al. 2013). Replication blocking lesions, such as thymine glycol, can change DNA structure and inhibit DNA replication by interfering with the activity of DNA polymerases (Basu et al. 1989; Kung and Bolton 1997). In addition, these DNA lesions can either slow down the activity of transcription complex, or completely blocks it, for instance, in the case of a SSB (Kathe, Shen, and Wallace 2004). As a result, DNA lesions can affect overall cell physiology, and may lead to mutation, cytotoxicity and ultimately disease. DNA damage response (DDR) pathway is a complex signaling network that has evolved to sense for DNA damages, activate cell signaling and repair DNA lesions. DNA damage are surveyed via DNA damage sensors, Mre11-Rad50-Nbs1 (MRN) complex, and replication protein A (RPA) and Rad9-Rad1-Hus1 (9-1-1) complex, that detects for DSBs and exposed regions of single stranded DNA respectively (Sulli, Di Micco, and d'Adda di Fagagna 2012). These sensor molecules then facilitate the recruitment of signal transducers in DDR pathway to the site of DNA damage. Signal transducers, including kinases ataxia telangiectasia mutated (ATM), ataxia telangiectasia mutated and Rad3-related (ATR) and DNAdependent protein kinase (DNA-PK), activate downstream mediator and effector 29 proteins to regulate cell cycle progression, DNA repair, transcriptional programs, cell senescence and cell death pathways (Chen et al. 2013; Zhou and Elledge 2000). For instance, DDR controls cell cycle progression via checkpoint effector kinases, Chk1 and Chk2, and induces cell arrest to inhibit DNA replication and mitosis when DNA damage such as a SSB or DSB occurs. This allows time for DNA repair to take place, preventing error during DNA synthesis and avoiding mitotic catastrophe (Stracker, Usui, and Petrini 2009; On et al. 2011). In another instance, p53, a master regulator protein of the DDR pathway can either be proapoptotic or promote cell-cycle arrest and DNA repair by varying its pattern and site of expression based on the severity of DNA damage (Chen et al. 2013). Hence, DDR can play pivotal roles in deciding ultimate cell fate following DNA damage. DNA repair mechanisms are critical to neutralize genotoxic DNA damage, and they form an important component of DDR pathway. Due to a wide diversity in DNA lesions, several DNA repair pathways are in place to correct different kinds of DNA damage. Mammalian cells uses five major DNA repair pathways: base excision repair (BER), nucleotide excision repair, mismatch repair and two DNA double strand break repair mechanisms (Altieri et al. 2008). These repair pathways are vital for preserving genome integrity of cells, such that impairment or deficiency in one or more DNA repair components are known to predispose individuals to cancer (Parshad et al. 1996; Peltomaki 2001). 1.3.2 Base excision repair removes damaged bases Base modifications are arguably the most common type of DNA damage upon ROS / RNS exposure. To counter with the detrimental effects of damaged nucleobases, BER, a highly coordinated DNA repair mechanism, has evolved to 30 repair non-bulky base lesions such as 8-OH-dG. BER occurs generally in five steps—removal of damaged base, incision of abasic sites, removal of ligation blocking sugar fragments, DNA polymerization to fill gaps, and finally, ligation of broken ends (Kim and Wilson 2012). BER is initiated by DNA glycosylases (e.g. 3-methyladenine-DNA glycosylase, NTH1, NEIL1, NEIL2 and OGG1) which cleave N-glycosidic bond to release damaged bases from the DNA backbone, leaving behind a highly mutagenic apurinic/ apyrimidinic (AP) site (Loeb and Preston 1986). AP sites are further processed by AP-endonucleases (e.g. APE1) or by binfunctional DNAglycosylases (containing AP-lyases activity) to incise DNA 5’ to the AP site. As a result, a SSB is generated, which has a 5’ end containing deoxyribose phosphate (dRP) or 3’ end containing blocking moiety that prevents direct ligation (Maynard et al. 2009; Caldecott 2008). For BER to complete, 3’-blocking terminus must be removed by AP endonucleases and polynucleotide kinase (Demple and Sung 2005; Wiederhold et al. 2004; Rasouli-Nia, Karimi-Busheri, and Weinfeld 2004), and 5’-dRP must be trimmed away by polymerase-β (Pol-β, containing 2deoxyribose-5-phosphate lyase activity) (Kim and Wilson 2012). If removal of blocking termini is not completed, it is not possible to ligate the broken ends together, leading to a physical SSB. To complete BER, missing nucleotide must first be replaced, and then the nick sealed. Missing nucleotides during BER can be filled via short- patch or long- patch BER pathways depending on the length of nucleotides incorporated. During shortpatch BER, Pol-β carries out DNA synthesis to fill single nucleotide gaps. In contrast, when cellular ATP level is low, multiple nucleotides ranging from two to 12 nucleotides are incorporated by Pol- and Pol- in long- patch BER, to generate 31 a flap intermediate that will be removed by flap endonuclease 1 (FEN1). Finally, the broken ends are ligated with DNA ligase I or ligase IIIα / XRCC1 complex to restore DNA into its undamaged state (Kim and Wilson 2012). 1.3.3 DNA damage can arise from BER intermediates Although BER is necessary to repair base lesions, other forms of DNA damages, namely AP - sites and strand breaks, can arise indirectly as BER intermediates. Evidence has demonstrated that the removal of 5’ dRP by Pol-β is rate limiting and that dRP-containing AP-intermediates can accumulate in the BER pathway (Srivastava et al. 1998; Nakamura and Swenberg 1999). When two or more unrepaired BER-induced SSBs are found close to each other on complementary strands, more harmful DSBs are generated. If replication blocking repair intermediates are not completely resolved before DNA synthesis, SSBs and DSBs can also indirectly arise due to replication fork arrest and replication fork collapse respectively. This may pose a challenge when actively dividing cells are exposed to DNA damaging agents (e.g. ROS / RNS), for instance, in injured tissue that undergo cell division for tissue regeneration. Interestingly, evidence show that multiple damaged sites with clustered base lesions carry extremely high risk of converting into DSBs. This has been attributed to incomplete BER-mediated repair in clustered base lesions which are extremely difficult to repair, leading to the formation of BER intermediates (Sedletska, Radicella, and Sage 2013; Kozmin et al. 2009; Bellon et al. 2009). Evidence has also suggested that when the formation of AP sites are prevented by inhibiting DNA glycosylases, generation of DSBs at the AP site is in turned decreased in yeast, bacteria and mammals (Sedletska, Radicella, and Sage 2013; Kozmin et al. 2009; Kidane, Murphy, and Sweasy 2014; 32 Ebrahimkhani et al. 2014b), further demonstrating that BER intermediates can contribute to more harmful DNA strand breaks. Taken together, DNA damage (in the form of DNA strand breaks) can be induced directly by ROS / RNS, as well as indirectly during the repair of ROS / RNS-induced base modifications. 1.3.4 DNA double strand break (DSBs) repair DSBs are one of the most toxic types of DNA damage that can be induced directly by potent ROS / RNS or indirectly by BER intermediates. Unlike other types of single stranded lesions, during DSBs, both DNA strands are damaged and there are no complementary templates that are intact for repair synthesis. When left unrepaired or improperly ligated, DSBs can block cell replication, lead to toxicity and introduce large scale genomic rearrangements. To prevent cytotoxicity induced by DSBs, homology-independent and homology dependent pathways are in place to repair these DNA lesions. Two dominant DSB repair pathways have been characterized, namely non-homologous end joining (NHEJ), a rapid DSB ligation pathway that does not require any sequence homology, and homologous recombination (HR), which requires extended homologous DNA sequences to retrieve genetic information lost during DSB formation (Jackson 2002). Both NHEJ and HR involve the recruitment and concerted activity of multiple DNA repair complexes that are required to process the severed DNA ends (Chapman, Taylor, and Boulton 2012) (Figure 1.5), although HR is far more elaborate than NHEJ. One study suggests that classical NHEJ can be completed in approximately 30 min, while HR can only be completed in 7 hours or longer (Mao et al. 2008a). The efficiency of NHEJ-mediated repair and its ability to operate in most cell cycles 33 are possibly some reasons why NHEJ is found to be the predominant DSB repair pathway in mammalian cells. NHEJ is initiated with the migration of Ku70 and Ku86 at the ends of DSB and the formation of a ring complex called Ku heterodimer (Ku70/Ku86). Ku heterodimer functions to protect DSBs from unwanted resection, and is critical in the recruitment of DNA-dependent protein kinase catalytic subunit (DNA-PKcs) that activate DNA repair proteins via Ser/Thr phosphorylation. Artemis may be involved as a nuclease that trims off secondary structures from the DSB ends (Ma, Schwarz, and Lieber 2005), while MRN complex (containing Mre11 nuclease) processes overhangs so that NHEJ can proceed. Finally, X-ray repair cross-complementing protein 4 (XRCC4) and ligase IV form a complex on DSB ends to mediate ligation (Figure 1.5) (Lieber 2010). In some situations, when Ku is deficient, DNA resection of DSBs occur revealing small patches of microhomologies (1-10 bp) that are involved in mediating ligation. In this case, an alternative NHEJ pathway, termed as alternative end joining (AEJ), takes place with ligation of DSBs via ligase III and XRCC3, followed by limited polymerization to fill gaps in between the ligated DNA. It has been thought that NHEJ, especially A-EJ, are error-prone processes that can lead to deletions due to end-processing of modified termini that cannot be directly ligated. NHEJ is also associated with the risk of translocation and gene rearrangements when distal ends are wrongly joined together (Lieber and Wilson 2010; Jackson 2002; Pfeiffer, Goedecke, and Obe 2000; Betermier, Bertrand, and Lopez 2014). In contrast to NHEJ, HR is considered an error-free DSB repair pathway that mostly operates during the late S to G2 phases of cell cycle. In somatic cells that 34 are not undergoing DNA synthesis, NHEJ may be an efficient repair pathway as compared to HR. However, evidence has been shown that NHEJ is highly errorprone during DNA synthesis, leading to high frequency of deletions and joining of non-fully complementary ends that result in mismatches, nicks and gap intermediates (Guirouilh-Barbat, Huck, and Lopez 2008). It has also been proposed that since replication forks are widely spaced apart, NHEJ may promote large scale genomic rearrangements during cell division process (Allen et al. 2011). The risk of NHEJ errors during DNA synthesis can explain why HR, a more conservative repair mechanism that retrieves lost genetic information using sister chromatids, is the dominant DSB repair pathway during mid S phase in mammalian cells (Mao et al. 2008b; Karanam et al. 2012). Classical HR begins with 5’ to 3’ resection of the broken DNA ends by MRN complex to generate 3’ overhangs that are then stabilized by RPA to prevent secondary structures. With the help of mediators such as Rad52 and Rad55/ Rad57, Rad51 is loaded onto the 3’ overhang to displace RPA and to form presynaptic filaments. Other proteins including Rad54 and Rhd54 help presynaptic filaments to conduct homology search and strand invasion (synapsis in undamaged sister chromatids or homologous chromosome), leading to heteroduplex DNA (D-loop) formation. During post synapsis, new DNA is synthesized by polymerases such as Pol-, Pol-ε and Pol-η to extend the invading 3’ end and fill the DSB gap (Maloisel, Fabre, and Gangloff 2008), followed by ligation of broken ends with DNA ligase I. The resulting heteroduplex contains branched structures in between the donor and recipient strands called Holliday junctions. Finally, BLM complex composed of BLM helicase, topoisomerase IIIα, RMI1, and RMI2 come into play to resolve the 35 Holliday junctions and to separate donor and recipient strands. In this process, BLM helicase unwinds DNA while topoisomerase IIIα induces topology changes in DNA by breaking and rejoining DNA, facilitating the formation of noncrossover products and completion of DSB correction. Alternatively, Holliday junctions can also be resolved via endonucleolytic cleavage with enzymes such as MUS81– EME1 in eukaryotes, leading to crossover formation (Pfeiffer, Goedecke, and Obe 2000; Krejci et al. 2012; Taylor and McGowan 2008) (Figure 1.5). Figure 1.5 Classical DNA double strand break repair pathways. (A) Non-homologous end joining (NHEJ) is a homology-independent DSB repair pathway. NHEJ is operational in all cell cycles, and is initiated by the binding of Ku70/Ku80 (also known as Ku70/Ku86) heterodimer at the two broken ends, to confer protection to the 36 broken ends and to facilitate recruitment of downstream proteins. Ku then recruits DNAPKcs, which together forms a holoenzyme that activates downstream proteins involved in signal transduction and DNA repair. Other DNA repair components involved are Artemis and MRN complex that play roles in DSB end processing to remove structures that inhibit ends joining. Finally DNA polymerase and XRCC4-Ligase IV complex come into play to catalyze limited polymerization (if any) and ligation of severed DNA ends. (B) Homologous recombination (HR) is a homology-dependent DSB repair pathway. Classical HR begins with 5’ to 3’ resection by MRN complex to generate 3’ overhangs stabilized by replication protein A (RPA). Rad52 interacts and colocalizes with Rad51, helping Rad51 displace RPA to form presynaptic filaments. Other proteins including Rad54 and Rhd54 help presynaptic filaments to conduct homology search and strand invasion leading to Dloop formation. New DNA is synthesized via DNA polymerase to fill the broken gap in the invading strand, and branched structures called Holliday junctions are formed as late intermediates in between the donor and recipient strands. Finally, dissolution of the Holliday junction or the donor and recipient strands take place to complete the HRmediated DSB correction via resolvases or BLM complex. 1.4 Studying significance of DNA repair pathway in vivo It is an undeniable fact that DNA repair pathways are essential for preservation of genomic stability and cellular integrity. Deficiency in key players of DNA repair pathways have been long found to be associated with, or evidently contribute to numerous diseases, including congenital defects, neurodegeneration, premature aging and cancer development (McKinnon 2009; Lehmann 2003; Sarasin and Stary 1997; Brosh and Bohr 2007). However, while DNA repair is indispensable for maintaining cellular integrity, excessive or uncontrolled activation of DNA repair pathways may counterintuitively lead to undesirable outcomes. For instance, it has been demonstrated that Alkyladenine DNA glycosylase responsible for the removal of damaged DNA bases during BER can lead to transient formation of abasic sites that, if not resolved in a timely manner, are replication and transcription blocking, highly mutagenic and may cause cell and even host lethality (Calvo et al. 2013; Yu et al. 2003; Ebrahimkhani et al. 2014a). In another instance, while HR is vital for DSB repair during DNA synthesis, it also carries the risk of misalignments that cause insertions, deletions, as well as loss of heterozygosity 37 (LOH), promoting cancer when such mutations occurs at tumour suppressor genes (Bishop and Schiestl 2001). Hence, to study the relationship between DNA damage and disease progression, it is also vital to gain further insights into the significance of DNA repair pathways, as protective or destructive mechanisms underlying disease outcome 1.4.1 Genetically engineered mouse that enable visualization of HR events Detailed phenotyping, and in-depth mechanistic understanding of DNA repair pathways that are activated during influenza infection, such as HR, can further our knowledge on the beneficial or detrimental contributions of DNA repair in disease outcome. However, investigation of the significance of DNA repair in vivo is limited by a lack of suitable animal models. Here, we propose to leverage on a Rosa26 Direct Repeat-Green Fluorescent Protein (RaDR-GFP) mouse which harbors a HR direct repeat substrate composed of two inactivated enhanced green fluorescent protein (EGFP) expressional cassettes inserted in tandem at the Rosa26 locus. When DSB occurs at the Rosa26 locus, full length EGFP can be reconstituted via HR-mediated repair, causing cells to fluoresce (Sukup-Jackson et al. 2014), thereby making it possible to isolate and monitor cells that have undergone HRmediated DSB repair. Hence, the RaDR-GFP mice serve as excellent tool for dissecting different HR mechanism and for examining cell fate, creating a better opportunity for understanding the roles of HR in influenza infection, and eventually, to facilitate the discovery of novel therapeutic targets. 38 1.5 Improving biodistribution of therapeutic agents Unfavorable biodistribution of therapeutic agents is a major hurdle for effective clinical interventions. A potential approach for improving the distribution of therapeutic agents at affected tissue is to use drug delivery systems that specifically target site of injury. It is now known that many influenza-related deaths are not a direct result of virus insult, but are attributed to dysregulated inflammatory responses and inflammation-induced lung injury. Taking into account observations from H5N1 and pandemic H1N1 outbreaks, whereby diffuse viral pneumonitis is the principle symptom observed among a huge proportion of hospitalized and intensive care patients (Bautista et al. 2010; Howard, Peiris, and Hayden 2011), this thesis aim to develop an approach for targeting pharmaceutical agents, such as antioxidants, specifically at the site of pulmonary inflammation and injury. 1.5.1 Targeting inflamed tissue with pH sensitive peptide Seeking characteristics that are unique to sites of inflammation, we turned our attention to tissue acidification. While most healthy organs are usually slightly alkaline (~pH 7.2-7.4) under physiological conditions, the interstitium of inflamed tissue has a much lower pH (as low as pH 5.8) (Punnia-Moorthy 1987; Martin and Jain 1993). In this thesis, we set out to develop a strategy for drug targeting by exploiting a ~36 amino acid pH-responsive peptide derived from bacteriorhodopsin of Halobacterium salinarum and related variants. This peptide, known as pH (low) insertion peptide™ (pHLIP) can change its conformation from an unstructured peptide at physiological pH to an α-helix that inserts into cell membrane at acidic microenvironments (Andreev et al. 2010). More importantly, pHLIP is capable of bidirectional delivery of compounds not only 39 to the extracellular surface of targeted cells, but also bring small, impermeable molecules intracellularly, making it a feasible approach for delivering polar drugs that are normally cell- impermeable (An et al. 2010). In addition, evidence has shown that pHLIP can be used to coat liposomes and other nanoparticles that allow bulk delivery and controlled release of compounds more specifically at targeted site as compared to nude liposomes or nanoparticles (Sosunov et al. 2013; Yao et al. 2013), making pHLIP a excellent approach for targeted drug delivery. Due to the unique biophysical property of pHLIP, it has consistently been shown to localize at regions of low extracellular pH such as tumors, known to be acidified by a locally high concentration of lactic acid generated via the Warberg phenomenon. Specifically, pHLIP is demonstrated to target both solid and metastatic tumors in vivo (Segala et al. 2009; Andreev, Engelman, and Reshetnyak 2009; Reshetnyak et al. 2010), and has also been shown to accumulate at inflamed tissue such as arthritic joints (Andreev et al. 2007). However, the potential utility of pHLIP had not been explored in the context of infectious diseases that involve acute inflammation. Given that pHLIP can target acidified tissues, we hypothesized that lung inflammation induced by influenza infection can cause pulmonary tissue to be acidic enough for the accumulation of pHLIP. There are several possible mechanisms whereby influenza-induced inflammation may result in lung acidification. During inflammation, polymorphonuclear cells (PMNs) can generate acidic ROS / RNS (e.g. HOCl and peroxynitrous acids) and lactic acid to promote local acidosis (Menkin 1960; Shepherd 1986; Comhair and Erzurum 2002). In the adaptive phase of infections, cytotoxic T cells can also release acidic contents in their cytotoxic granules to lower surrounding pH (Henkart et al. 1987). Hence, we 40 hypothesized that influenza infected mouse lungs are acidified and can be specifically targeted by pHLIP in order to improve biodistribution and efficacy of therapeutic agents conjugated to the peptide. 1.6 Thesis aims A substantial body of evidence suggests that inflammatory responses play important roles in driving disease progression during influenza infection. This dissertation addresses fundamental questions primarily on pathology and host responses, and in part on the treatment of influenza-induced lung inflammation. Oxidative stress elevated during influenza infection carries with it potential risk for damage to essential biomolecules. One goal of this study is to further our understanding of the molecular pathology associated with influenza infection and influenza-induced inflammation. We are particularly interested in the impact of infection on host DNA, since the genome is vital for building and maintaining an organism. Another goal is to help in developing a mouse model to study HR, a major DNA repair pathway, in lungs. This work sets a foundation for future investigations on potential molecular factors underlying disease mechanisms and disease susceptibility for severely infected individuals. The thesis also aims to investigate the potential of a novel strategy that can be exploited to improve the biodistribution, and hence the effects, of therapeutic treatments. We evaluated the localization properties of a low-pH targeting peptide, pHLIP, in a primary influenza pneumonia model. This study not only enables us to determine if pHLIP can specifically distribute small molecules to the site of lung injury, but also allows us to explore the potential of using inflammation-induced tissue acidity as a targeting feature for drug delivery vehicles. 41 Chapter 2 Methods and materials 2.1 Materials 2.1.1 Media, chemicals and reagents Table 2.1 Commercial sources of media and reagents Experiment Cell/virus culture Company Invitrogen (Gibco) Reagent/Kit Dulbecoo’s Modified Eagle Medium (DMEM) Minimum essential media (MEM) Fraction V Bovine Serum Albumin (BSA) Penicillin-streptomycin-Glutamine (100x) Trypsin EDTA Phosphate buffered saline (PBS) 10x Fetal bovine serum (FBS) Avicel TPCK-treated trypsin (Bovine pancreas) Target retrieval solution 10x 4',6-Diamidino-2-Phenylindole (DAPI) ProLong Gold Antifade Mountant with DAPI L-1-Tosylamide-2-phenylethyl chloromethyl ketone treated trypsin (TPCK-Trypsin) Proteinase K See below 2-N-morpholino-8-dibenzothiophenylchromen-4-one (NU7441) Xanthine oxidase Fluorometric assay kit 8-hydroxy-2’-deoxyguanosine/ DNA/RNA EIA kit Mouse TNF-alpha Quantikine kit Bradford assay DC protein estimation assay 30% polyacrylamide 20% Sodium dodecyl sulfate (SDS) 1M Tris-HCl pH8 1.5M Tris-HCl pH6.8 Ammonium persulfate (APS) Tetramethylethylenediamine (TEMED) PVDF/Nitrocellulose membrane Tween-20 Phosphate buffered saline (PBS) 10x Low fat milk WesternBright ECL HRP substrate Amersham ECL Prime Vivantis Hyclone FMC Sigma Immunofluoresence DAKO Invitrogen Sigma Aldrich Chemicals Promega Antibodies Axon Medchem Biochemical assays Cayman Western blotting R&D systems Biorad Vivantis Anlene Advansta GE lifesciences 42 Table 2.2 Formulations for solutions and buffers Solutions 2x Laemmli buffer Formulation (reagents from Sigma Aldrich) 65.8 mM Tris-HCl, pH 6.8, 2.1% SDS, 26.3% (w/v) glycerol, 0.1M DL-Dithiothreitol (DTT) SDS-PAGE Running buffer 25 mM Tris, 192 mM glycine, 0.1% SDS, pH 8.3 SDS-PAGE Transfer buffer 25 mM Tris, 192 mM glycine, 20% (v/v) ethanol, pH 8.3 Citric acid buffer 10mM Citric Acid, pH 6.0 2.1.2 Cell cultures Madin-Darby canine kidney (MDCK) cells are a canine kidney epithelial cell line used for plaque titration of influenza containing samples (plaque assay). MDCK cells were maintained in MEM medium supplemented with 5 – 10 % FBS and grown in a 37 ⁰C humidified incubator with 5 % carbon dioxide (CO2). Cells were split twice a week and up to 10 passages were used for plaque assay. Adenocarcinomic human alveolar basal epithelial cells (A549) and baby hamster kidney (BHK) cells used for testing the pH sensitivity of pH (Low) Insertion peptide (pHLIP) and pHLIP variant peptides were maintained in DMEM supplemented with 10 % FBS and 1 % penicillin-streptomycin-glutamine mixture. 43 2.1.3 List of antibodies for western blot and immunofluorescence Table 2.3 Sources and clones of antibodies used S/N 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 Target (Source) Western blot Immunofluorescence (clones) clones HA (Sino biological) Clone ID 2 (#86001-RM01) Gamma-H2AX (Millipore) JBW301 Gamma-H2AX (Cell 20E3 signaling) CCSP (Santa Cruz) T-18 SPC (Santa Cruz) M-20 NS1 (Santa Cruz) NS1-23-1 Ku70 (Cell signaling) D10A7 Ku86 (Santa Cruz) M-20 Rad51(Santa Cruz) H-92 PCNA(Santa Cruz) C-20 PARP-1 (Abcam) Ab6079 PARP-1 (BD Transduction) 42/PARP (#611039) Cleaved PARP-1 (Abcam) E51 PAR (Trevigen) 4336-BPC-100 Cleaved caspase 3 Asp175 (#9661) Asp175 (#ab9661) (Abcam) F4/80 (conjugated, CI:A3-1 Biolegends) Ref-1 (Santa Cruz) C-4 Ki-67 (Dako) TEC-3 CD3 (conjugated, 17-A2 Biolegends) Podoplanin (R&D systems) #AF3244 Fibronectin (Sigma # F3648 Aldrich) Peroxiredoxin-6 (Abcam) #ab59543 Secondary antibodies (Donkey immunofluorescence, conjugated anti-primary host IgG (H+L)) for with Alexafluor-488, Alexafluor-546 and Alexafluor-647, were purchased from Life technologies. Secondary antibodies for western blot, conjugated with horse radish peroxidase (HRP), were purchased from DAKO. 44 2.1.4 List of antibodies for flow cytometry Two panels of antibodies were used for characterizing leukocytes in bronchoalveolar lavage fluid (BALF) cells. Table 2.4 Panel 1 antibodies for myeloid cells S/N 1 2 3 4 5 6 Primary target CD45 Siglec F CD11b CD11c MHC-II GR-1 Conjugate APC PE PE-Cy7 Pacific Blue FITC PerCP Cy5.5 Company Miltenyl Biotec BD Pharmingen eBioscience Biolegend Miltenyl Biotec eBioscience Table 2.5 Panel 2 antibodies for lymphoid cells S/N 1 2 3 4 5 2.1.5 Primary target CD45 CD3 CD4 CD8a CD19 Conjugate PE-Cy7 APC PerCP-Cy5.5 Pacific Blue FITC Company eBioscience Miltenyl Biotec Biolegend BD Pharmingen Miltenyl Biotec Source of influenza viruses H1N1 Influenza A/Puerto Rico/8/34 (PR8) virus was propagated in chick embryos by Mrs Phoon Meng Chee (NUS, Department of Microbiology). PR8 propagated in two eggs, egg 12 and egg 15, were titrated with plaque assay and used for the entire project. The sublethal titer for PR8 derived from egg 12 was 30 plaque forming units (PFU), while the sublethal titer for PR8 derived from egg 15 was 12 - 15 PFU. Both sublethal doses were tested on C57Bl/6 mice, resulting in comparable disease kinetics in terms of weight changes and lung pathology. 45 2.1.6 pHLIP and pHLIP variant peptide sequences and conjugation pHLIP peptides and pHLIP variant peptides were initially obtained from Dr. Damien Thévenin, (Ph.D., Assistant Professor of Chemistry, Lehigh University), and subsequently purchased from Anaspec. The sequences of each peptide variant are shown below: Table 2.6 Sequences and molecular weights of pHLIP and peptide variants Peptide Peptide sequence (Hylite-Fluor 647conjugates) pHLIP Leu26Gly K-pHLIP D3KpHLIP EQNPIYWARYADWLFTTPLLLLDLALLVDADEGTG EQNPIYWARYADWLFTTPLLLLDGALLVDADEGTG EQNPIYWARYAKWLFTTPLLLLKLALLVDADEGTG DDDEQNPIYWARYAKWLFTTPLLLLKLALLVDADEGTG Mass (MW) 4980.8 4924.7 5007 5352.3 The original pHLIP peptide obtained from Dr. Damien Thévenin was synthesized with a cysteine residue at its N-terminus by solid-phase peptide synthesis and purified by reverse-phase HPLC at the W.M. Keck Foundation Biotechnology Resource Laboratory (Yale University). Then, fluorescent dyes were conjugated to the N-terminal cysteine of pHLIP. Briefly, a solution of pHLIP peptide (1 µmol) dissolved in 1:1:3:5 of DMF, DMSO, methanol, and aqueous NH4HCO3 buffer (200 mM; pH 8) was added to 1 eq. of Alexa Fluor® 647 C2 maleimide (Invitrogen) in 50 µL DMSO. This mixture was stirred in the dark for 10 min at room temperature. The mixture (which was reduced) was then added to an oxidizing solution of K3Fe(III)CN6 (6 mg, 18.2 µmol, 23.1 eq.) in 60 µl of aqueous NH4HCO3 buffer (200 mM; pH 8). The reaction mixture was stirred at room temperature for 3 h. Desired peptide-conjugates was isolated via reverse-phase HPLC (Hewlett Packard Zorbax semi-prep 9.4 x 250 mm SB-C18 column; flow rate: 2 mL/min; phase A: water + 0.01% TFA; phase B: acetonitrile + 0.01% TFA; gradient: 70 min from 99:1 A/B to 1:99 A/B). 46 In subsequent experiments, Hylite Fluor 647-conjugated pHLIP, Leu26Gly, KpHLIP and D3K-pHLIP was purchased from Anaspec. The purity was ≥85% as confirmed by analytical HPLC. 2.2 Methods and protocols 2.2.1 Infection of mice and tissue collection Animal protocols were conducted in strict accordance with National Advisory Committee for laboratory Animal Research (NACLAR) Guidelines (Guidelines on the Care and Use of Animals for Scientific Purposes) in facilities licensed by the Agri-Food and Veterinary Authority of Singapore (AVA), the regulatory body of the Singapore Animals and Birds Act. The protocol was approved by the Institutional Animal Care and Use Committee (IACUC), National University of Singapore (Permit Number: IACUC 117/10). 8-12 weeks old C57Bl6 mice were anaesthetized with 75 mg/kg of ketamine and 1 mg/kg of medatomidine, and then infected with a sublethal dose of PR8 by intratracheal instillation. In contrast, uninfected controls were instilled with same volume of sterile PBS, the vehicle used for PR8 dilution. Uninfected mice treated with PBS or saline is a widely accepted control for influenza animal models, which elicits minimal to no immune responses and is comparable to other controls such as mice treated with UV-inactivated influenza virus (Buchweitz, Harkema, and Kaminski 2007; Yasui, Kiyoshima, and Hori 2004; Sundar et al. 2015; Monteiro, Harvey, and Trinchieri 1998). Mice instilled with PR8 or vehicle control were then revived with 1 mg/kg of atipamezole. Two batches of lungs from PR8 infected mice were harvested at various time points up to 17 days post infection (dpi). Uninfected controls were euthanized and their organs were harvested on 3 dpi, 5 dpi, 9 dpi 47 and 13 dpi (two samples were collected each day). Left lung was fixed in 10% neutral buffered formalin and paraffin embedded for histology, while right lung were snap frozen in liquid nitrogen or lavaged with 1ml ice-cold PBS to collect BALF. 2.2.2 Lung homogenization and virus titration Frozen apical and cardiac lobes were homogenized with 300 µl of PBS with Qiagen’s Stainless steel beads and Tissuelyser, at maximum oscillation speed for 2 min at 4⁰C. Lung homogenate was spun down at 3000 rcf for 10 min at 4 ⁰C. The supernatant was aliquoted and stored at -80 ⁰C. To perform virus titration, MDCK cells were seeded at a density of 2x105 cells per well in a 24 well plate overnight in MEM containing 10% FBS. Next day, cells were washed twice with PBS. 100 µl of 10-fold serially diluted lung homogenate were then inoculated onto each well of MDCK cells (approximately 80% confluent) and incubated at 37⁰C for 1 h. Infected MDCK cells were then cultured in serum free MEM (Gibco) with 1.2% Avicel, 2 mg/ml TPCK trypsin (Sigma) and 0.2% Fraction V BSA (Gibco). After 48-72 h, cells were then fixed with formalin and stained with 1% crystal violet Plaques forming units (PFU) enumerated were normalized with protein concentration of lung homogenate estimated with Bradford assay (Biorad). 2.2.3 Lung histology, infiltration index calculation and pathology analysis Transverse sections of paraffin embedded lungs (5 µm) were stained with Haematoxylin and Eosin (H&E) staining based on previous reference with minor modifications (Fischer et al. 2008). Stained sections were scanned using Mirax Midi slide scanner (Carl Zeiss) and whole lung images were analyzed using Matlab (The Math Works Inc., Natick, Massachusetts) to calculate infiltration index. Infiltration index, defined as the percentage of dense areas (with multiple layers of 48 nucleated cells) to total alveolar area in H&E stained lung sections, was calculated based on previous publication (Yin et al. 2013). Briefly, original RGB image of H&E stained tissue section was converted to grayscale. One intensity threshold was set to exclude large empty spaces, including bronchi, trachea and large blood vessels, from total section area to obtain total alveolar region. Another intensity threshold was set to segment regions of infiltrated lung parenchyma that are darkly stained with H&E from uninfiltrated regions that possess lightly stained lung parenchyma. The percentage was computed from regions of infiltration and total alveolar region. Pathology analysis of H&E stained sections were performed by an independent pathologist, Dr Tan Kong Bing (Assistant Professor TAN Kong Bing, MBBS, FRCPA, FRCpath). Sections from a total of 3 - 5 mice per time points were analyzed. 2.2.4 Measurement of ROS markers and TNF-α XO quantification was performed with diluted lung homogenate using Xanthine oxidase Fluorometric assay kit (Cayman) based on manufacturer’s protocol. To measure oxidative damage to nucleic acids, BALF was collected and centrifuged, and the supernatant was collected and analyzed with DNA/RNA Oxidative Damage EIA Kit (Cayman) that measures the levels of free 8-hydroxydeoxyguanosine (8-OH-deoxyG), 8-hydroxyguanosine (8-OH-G) and their nucleobases. To measure the levels of soluble TNF-α, BALF was analyzed with the Mouse TNF-alpha Quantikine ELISA Kit (R&D systems) based on manufacturer’s protocol. 49 2.2.5 Western blotting Frozen middle and inferior lobes (right lungs) were homogenized with 2x Laemmli sample buffer with DTT, boiled for 10 min to denature protein, and then spun down at 13,000 rcf for 10 min. Protein concentration in each sample was estimated with DC Protein reagent (Biorad) based on manufacturer’s protocol. Same amount of protein was loaded onto 6-15% SDS-PAGE for each set of samples. Antibodies used for probing specific antigens included anti-Haemagglutin (HA; Sinobiological Inc.), anti-H2AX (Millipore), anti-Rad51, anti-Ku86 (Santa Cruz), anti-PCNA (Santa Cruz), anti-Ku70 (Cell Signaling), anti-PARP-1 (ab6079 Abcam), antiPARP-1 (cleaved p25) (Millipore), anti-PAR (Trevigen), anti-cleaved caspase 3 (Cell signaling) and anti-β actin (Sigma) as listed in Table 2.3. Each blot contained samples from different mice. Blots were exposed on film and band intensities were quantified using Thermo Scientific myImageAnalysis v1.1. All bands’ intensities were normalized with uninfected controls and normalized with housekeeping loading controls β-actin. 2.2.6 Flow cytometry of immune cells BALF cells (from right lung lavaged with 1 ml PBS) were pelleted and incubated with 1 ml ammonium-chloride-potassium (ACK) lysis buffer (Life technologies) for 5 min at room temperature. Cells were then stained with fluorophore conjugated antibodies to identify leukocytes (CD45+), alveolar macrophages (Siglec F+/CD11c+), eosinophils (Siglec F+/CD11c-), neutrophils (Siglec F- /GR1+/CD11b+), CD8+ T cells (CD3+/CD8a+) and CD4+T cells (CD3+/CD4+) in staining buffer (PBS containing 1% BSA) for 30 min at room temperature (referenced from (Zaynagetdinov et al. 2013; Han 2013 ; Buchweitz et al. 2007)). 50 All antibodies listed above were purchased from BD Pharmingen, eBiosciences or Miltenyi biotech. Stained cells were washed and analyzed with BD LSRFortessa (BD Bioscience). Flow cytometry plots were analyzed with FlowJo (Tree star). 2.2.7 Immunofluorescence assay Paraffin-embedded lung sections were dewaxed and hydrated. Antigen retrieval was performed with three methods, depending on the target used: Table 2.7 Antigen retrieval methods Antigen retrieval Protocol method Proteinase K digestion Incubate hydrated tissue sections with working proteinase K solution (20 µg/ml in TE Buffer, pH 8.0) for 30 min at 37⁰C Hot Critic Acid buffer Boil tissue sections in hot citric acid buffer (10 mM Citric Acid, pH 6.0) for 30 min in a microwave Dako Antigen retrieval Heat tissue sections in sub-boiling Dako antigen solution retrieval solution (pre-boiled) for 30 min at low microwave heat Sections were then blocked and permeabilized with 10% Donkey serum in PBS with 0.3 % Triton-x 100 for 1 h at room temperature. They were then incubated overnight at 4 ⁰C with 20 µg/ml of anti-H2AX (Cell Signaling), 1 µg/ml of anti-CC10 (or CCSP; Santa-Cruz), 1 µg/ml of anti-pro-SPC (Santa-Cruz), 1 µg/ml of anti-NS1 (Santa-Cruz), 1 µg/ml anti-Pdpn (R&D systems) or 1 µg/ml anti-Prdx6 (Abcam) in staining buffer (5% donkey serum / 3% BSA and 0.3 % Triton-x 100 in PBS). Sections were washed and incubated with 5 µg/ml of Alexafluor dyes-conjugated secondary antibodies (Molecular probes) for 1 h at room temperature on the following day, and then further stained with DAPI (Santa Cruz/ Life technologies) for 15 min. Sections were washed and mounted with ProLong gold antifade reagent (Life technologies). 51 To co-stain for Ki-67 and H2AX, lung sections were antigen-retrieved with DAKO antigen retrieval solution and first stained with 1 µg/ml of rat anti-Ki67 (DAKO) for 4 h at room temperature, followed by counter staining with anti-rat secondary antibodies conjugated to Alexafluor-546 for 1 h. The sections were then co-stained with rabbit anti-H2AX overnight followed by incubation with anti-rabbit secondary antibodies conjugated with Alexafluor-647 (see method described in previous paragraph). To stain cryosections, sections (10 µm) were fixed in 4% PFA for 15 min at room temperature and stained with 1 µg/ml of anti-CD3 (eBioscience) and 20 µg/ml of anti-H2AX (Cell Signaling) based on the protocol described above, except that incubation of primary antibodies were shortened to 1 h at room temperature. This is followed by staining with secondary antibodies based on previously described protocol. 2.2.8 Microscopy All sections were either imaged at 20x magnification with Mirax Midi slide scanner or at 40x magnification with Zeiss LSM 700 confocal microscope (Carl Zeiss) at a thickness of 3 µm. Bronchial epithelium were identified by positive Club cell secretary protein (CCSP) staining and pseudostratified columnar tissue structure. Alveolar type I (AEI) cells were identified by positive Pdpn immunostaining and alveolar type II (AEII) cells were identified by positive pro-surfactant protein C (SPC) immunostaining. To determine the frequency of H2AX positive bronchiolar epithelium, all bronchi and bronchioles were captured from each lung section. In order to collect images for alveolar parenchyma (CCSP negative region), 10 random regions were captured per lung section. 52 For peptide delivery assessment, wherein fluorophore-labelled peptides were injected in mice, tissue sections were first imaged with Carl Zeiss scanning microscope for fluorescence images. The sections were then washed with PBS thrice and counter-stained with H&E before they were re-imaged. 2.2.9 Manual and semi-automated quantification of H2AX foci Ten images of bronchioles, pro-SPC positive cells and lung parenchyma were selected and counted for each mouse in a blinded manner. To quantify nuclei in the lung parenchyma, DAPI stained nuclei were counted using Imaris version 7.6.5. At least 1000 (to > 3000) cells in the lung parenchyma were counted for each mouse. Nuclei of bronchial epithelium and pro-SPC positive cells were counted manually. Bronchioles were first identified by the presence of CCSP staining in a lumen lined by pseudostratified columnar epithelium. The number of pseudostratified columnar cells in the bronchioles was then counted manually regardless of CCSP expression. Cellular debris and immune cells in the lumen of bronchioles after infection were not counted. At least 400 bronchiolar epithelial cells were counted for each mouse. Pro-SPC positive cells were quantified by counting nuclei surrounded by pro-SPC staining. More than 100 cells were counted for most mice except for 5 mice (53 90 cells were counted) because there were less Pro-SPC positive cells in the captured images. To prevent bias, fluorescence channel for H2AX was switched off while manually counting the number of nuclei. Counted nuclei were labelled using manual spot function on Imaris so as to identify counted cells. For DSB analysis, cells harboring 5 or more foci were considered positive for H2AX. Cells 53 with pan-nuclear H2AX, i.e. cells possessing nuclei that were uniformly stained positive for H2AX, were quantified separately. To determine the relationship between cell division and DSB formation, 15 random images were taken for lung sections co-stained with Ki-67 and H2AX (total lung area analyzed = 0.4 mm2). The number of nuclear-Ki-67 positive, H2AX positive (≥ 5 foci), and Ki-67 / H2Ax dual positive cells were enumerated manually for each image. The frequency of H2Ax+ cells among all Ki-67 positive cells was quantified. The total number of Ki-67+ and H2AX+ cells counted in 0.4mm2 lung area varied at different time points and are reported below. Table 2.8 Mean number of Ki-67+ and H2AX+ cells counted in 15 40x magnified images Dpi Mean Std. Deviation Total Ki-67 cells counted Uninf. 5 dpi 42 109 12 33 9 dpi 489 132 13 dpi 362 121 Dpi Mean Std. Deviation Total H2AX cells counted Uninf. 5 dpi 13 112 5 81 9 dpi 172 29 13 dpi 171 69 2.2.10 Terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) and quantification Paraffin sections were dewaxed and treated with Proteinase K (20 ug/ml in TE Buffer, pH 8.0) for 30 min at 37⁰C, washed and then stained with In Situ Cell Death Fluorescein Detection Kit (Roche). Briefly, tissue sections were washed two times in PBS and incubated with TUNEL labeling mixture for 1 h at 37⁰C. Lastly, sections 54 were washed, stained with DAPI and mounted with ProLong gold antifade (Life Technologies). TUNEL stained sections were scanned with Mirax Midi slide scanner (Carl Zeiss). Ten randomly chosen images (at 40x digital magnification) were cropped from scanned lung sections and saved as .tif files. Selection was made nonbiased by switching off the TUNEL (FITC) channel on Panoramic viewer (3DHISTECH) used to access Mirax midi files. The number of TUNEL positive nuclei on each image was then quantified automatically using Imaris by segmenting DAPI channel to enumerate the number of nuclei. This is followed with quantification of nuclei overlapped with FITC channel (TUNEL positive cells). 2.2.11 Microarray data analysis Whole lung microarray data was retrieved from NCBI Gene Expression Omnibus (GEO) under the accession number GSE42639. Methods of data collection were published by Brandes et al. (2013) (Brandes et al. 2013). Briefly, Brandes et al. infected 5 - 8 week old C57Bl/6 mice with various doses of pathogenic PR8 strain at the indicated times post infection. Uninfected controls were administered physiological saline (Saline controls). Lung transcriptome was evaluated with Illumina MouseWG-6 v2.0 expression beadchip. Published series matrix table contains final values that were normalised by quantile normalisation using the GeneSpring GX 10.0 analysis platform (Agilent), log2 transformed and median baseline transformed. Heat map of expressional changes in genes involved in HR and NHEJ were created using Microsoft word excel. The data from a sublethal infection dose of 0.6LD50 was chosen for analysis as this dosage (high sublethal 55 dose) manifests with disease kinetics that most closely matches that of our Influenza A murine model. 2.2.12 NU7441 treatment of infected mice C57Bl/6 mice were infected with sublethal dose of PR8. 2-N-morpholino-8dibenzothiophenyl-chromen-4-one (NU7441), a small molecule ATP-competitive inhibitor of DNA-PK (purchased commercially from AXON), was reconstituted at 1 mg/ml with 40% polyethylene glycol (PEG) in sterile physiological saline. Reconstituted NU7441 was homogenized with a sonicator (Osim) for 5 min at room temperature. In our study, mice were injected with two or three doses of 10 mg/kg of NU7441 intraperitoneally on 9 dpi and 13 dpi, every 4 h apart. Untreated mice were injected with vehicle control, which is 40% PEG/saline. After the last dose of NU7441, mice lungs were harvested after 1 h post injection for further evaluation. 2.2.13 RaDR mice and single cell suspension preparation RaDR-GFP mice were generated in C57Bl/6 background based on the description in previous publication (Sukup-Jackson et al. 2014). Adult mice (3–4 months old) were housed and handled in accordance to protocol approved by the Institutional Animal Care and Use Committee [MIT CAC (IACUC)]. Animals were sacrificed with carbon dioxide asphyxiation and organs such as pancreas and spleen were excised. Excised organs were washed with sterile HBSS (Gibco) and each placed in 10 ml of DMEM containing 2 mg/ml collagenase D (Sigma) and 40 U/ml DNase I (Invitrogen). Organs were dissociated in gentleMACS C tubes (Miltenyi Biotec) with gentleMACS Dissociator using preloaded settings for the individual organs except pancreas which were disintegrated with the settings for brains. Dissociated organs were then placed on a rocking shaker for 30 min in 37⁰C, and then put 56 through a 70 µm mesh cell strainer. Single cell suspension was washed with HBSS and used for downstream FACS sorting. 2.2.14 RaDR cells RNA extraction and cDNA conversion Approximately a million embryonic stem (ES) cells or 1000 cells (EGFP positive or EGFP negative) were sorted from RaDR mouse pancreas using MoFlo (Beckman Coulter) or BD FACSAria (BD biosciences) directly into 200 µl of TRIzol® (Lifetechnologies). These are then topped up to 1 ml of TRIzol® and stored at -80⁰ for future RNA extractions. Total RNA was extracted with column based purification using the RNeasy mini kit (Qiagen). Briefly, 200 µl of chloroform was added to Trizol-lysed cells and spun down at 12,000g for 15 min at 4⁰C. The aqueous layer formed was added to 500 µl of ice-cold isopropanol and then applied to an RNeasy column. The column was washed, as per manufacturer’s protocol, once with RW1 and twice with RPE buffer, followed by elution with 30 µl of RNasefree water. Total RNA was converted to cDNA with SuperScript™ III First-Strand Synthesis System for RT-PCR (Lifetechnologies) with both random hexamers and oligo(dT). Briefly, 1 µl of oligo(dT) and 1 µl of random hexamers provided by the manufacturer were added to 1 µl of 10 mM dNTP mix (10 nmol) and 500-2000 ng of total RNA. The mixture (topped up to 10 µl with RNase-free water) was then incubated at 65⁰C for 5 min and placed on ice for at least 1 min. After adding the reverse transcriptase master mix based on given manufacturer’s protocol, the mixture was incubated at 25 ⁰C for 10 min, 50 ⁰C for 50 min, and 85 ⁰C for 5 min and then placed on ice. 1 µl of E.coli RNase H was added and the mixture was incubated at 37⁰C for 20 min to remove RNA-cDNA duplex before proceeding with PCR. 57 2.2.15 Direct PCR analysis using RNA transcripts cDNA was diluted 10 times with nuclease-free water. Using 5-10 µl of the diluted cDNA, PCR detection of full-length EGFP sequences was performed by using primers A FL FOR‘ ATTCGC CACCATGGTGAGCAAGGGC’ and C FL REV ‘GATATCAAGCTTACTTGTACAGCTC’ using Platinum Taq DNA Polymerase (LifeTechnologies). In the presence of 0.2 µM of each of the primers and enzyme mix based on manufacturer’s protocol, cDNA was denatured at 94⁰C for 3 min, and then incubated for 40 cycles at 94 ⁰C for 45 sec, 56 ⁰C for 45 sec and 72 ⁰C for 1.5 min. The PCR mixes were then incubated at 72⁰C for a final 5 min and placed on ice. In order to detect Δ5egfp and Δ3egfp, 2 primer sets were used in one reaction. They are E D5 FOR2 ‘ACAGGGTAACTAGCTGGATCC’ and F D5 INT REV ‘TGCTTCATGTGG TCGGGGTAGCGG’ that detects for Δ5egfp and G D3 INT FOR ‘TTCTTCAAGTCC GCCATGCCCGAA’ and H D3 REV2 ‘AGATGCTGAGGTACCGGATCCTAT’ that detects for Δ3egfp. Each reaction contained 0.2 µM of the primers. PCR reactions were incubated at 94⁰C for 3 min, and then at 94 ⁰C for 45 sec, 55 ⁰C for 30 sec and 72 ⁰C for 1 min 10 sec for 40 cycles. Finally, they were incubated at 72 ⁰C for a final 5 min and placed on ice. Please see Table 2.9 for a summary of primer sequences and Table 2.11 for PCR cycling conditions. 58 Table 2.9 PCR primers to specifically amplify full length EGFP, Δ3egfp, and Δ5egfp. Gene Direct PCR/Internal PCR primers Primer sequence Full length EGFP A FL FOR ATT CGC CAC CAT GGT GAG CAA GGG C C FL Rev GAT ATC AAG CTT ACT TGT ACA GCT C E D5 FOR2 ACA GGG TAA CTA GCT GGA TCC F D5 INT REV TGC TTC ATG TGG TCG GGG TAG CGG G D3 INT FOR TTC TTC AAG TCC GCC ATG CCC GAA H D3 REV2 AGA TGC TGA GGT ACC GGA TCC TAT Δ5egfp Δ3egfp 2.2.16 Nested PCR analysis for Full-length EGFP Using a limited number of pancreatic cells (e.g. 1000 cells per sample), direct PCR is often not sensitive enough to detect full-length EGFP. In order to increase the chances of detecting full length EGFP, we performed a nested PCR. External PCR primers are placed upstream the EGFP gene downstream of the predicted promoter region (forward primer) and downstream in the poly A site (reverse primer). 0.2 µM of primers BPEF3 ‘CTGACTGACCGCGTTACTCC’ and NEST Rev ‘TAGGCAGCCTGCACCTGAG’ was added to Platinum Taq DNA polymerase mix with 5-10 µl of 10x diluted cDNA following manufacturer’s protocol (Table 2.10). External PCR mixes were incubated at 94 ⁰C for 3 min, then at 94⁰C for 45 sec, 58⁰C for 30 sec and 72⁰C for 1 min 10 sec for 40 cycles, and finally at 72⁰C for a final 5 min and placed on ice (Table 2.11). External PCR products are purified using MinElute PCR Purification Kit (Qiagen) and eluted with the same volume of EB buffer. Five microliters of purified amplicons was used for subsequent full 59 length EGFP PCR as described in previous section. All final PCR products were analyzed by resolving on a 1.5% agarose gel and the expected band sizes are listed in Table 2.11. Table 2.10 External PCR primers designed to anneal upstream and downstream of the EGFP coding sequence. External PCR primers Primer sequence BPEF3 CTG ACT GAC CGC GTT ACT CC Nest Rev TAG GCA GCC TGC ACC TGA G Table 2.11 Thermal cycler conditions and product sizes. Gene Primer (Sense) Primer (Antisense) Size (bp) PCR conditions Full length EGFP ATT CGC CAC CAT GGT GAG CAA GGG C GAT ATC AAG CTT ACT TGT ACA GCT C 740 94 ⁰C, 3 min 94 ⁰C, 45 sec 56⁰C, 45 sec 72⁰C, 1:30 min Repeat from step 2, 39x 72 ⁰C, 5 min 4 ⁰C, infinity Δ5egfp Δ3egfp ACA GGG TAA CTA GCT GGA TCC TTC TTC AAG TCC GCC ATG CCC GAA TGC TTC ATG TGG TCG GGG TAG CGG 415 AGA TGC TGA GGT ACC GGA TCC TAT 250 94 ⁰C, 3 min 94 ⁰C, 45 sec 55 ⁰C, 30 sec 72⁰C ,1:10min Repeat from step 2, 39x 72 ⁰C, 5 min 60 4 ⁰C, infinity External PCR CTG ACT GAC CGC GTT ACT CC TAG GCA GCC TGC ACC TGA G 837 94 ⁰C, 3 min (predicted) 94 ⁰C, 45sec 58 ⁰C, 45sec 72 ⁰C, 1:10min Repeat from step 2, 39x 72 ⁰C, 5 min 4 ⁰C, infinity 2.2.17 Single cell nested PCR analysis (DNA PCR) Single cells from RaDR mouse spleen are sorted individually into 5 µl of lysis buffer (400 ng/µl of proteinase K and 17 µM SDS in nuclease-free water) with BD FACSAria. Similarly, a single colony of RaDR-EGFP ES cells was also picked up with a pipette tip and lysed with the same lysis buffer. The cell lysate was then freeze-thawed once at -80 ⁰C to release cellular contents, and added to a total volume of 50 µl Platinum Taq DNA Polymerase (LifeTechnologies) mix with 0.2 µM of primers BPEF3 and NEST Rev (Table 2.10). External PCR was performed according to conditions described in Table 2.11. Five microliters of external PCR product was then transferred to FL-EGFP internal PCR. Two microliters of external PCR product was used in Δ3 Δ5egfp internal PCR following previously described protocol (Table 2.9, Table 2.11). 61 2.2.18 In vitro experiment with pHLIP Chamber slides were prepared with A549 cells (~5x104 cells per well), cultured for 24 h, and rinsed twice with PBS adjusted to pH 5, 6, 7.4 and 8. Following incubation with 8 µM of 5-FAM-conjugated pHLIP in PBS with appropriate pH (30 min, room temperature, in the dark), cells were washed with the appropriate pH-adjusted PBS and viewed under a fluorescence microscope. In later experiments, BHK cells were incubated with 8µM of hylite fluor-conjugated pHLIP variants (pHLIP, Leu26Gly, KpHLIP, D3K-pHLIP) in DMEM (containing 10% FBS) adjusted to pH 5, 6, 7.4 and 8 for 30 min at room temperature. 2.2.19 Peptide injection in mice Mice were injected intraperitoneally with 10 nmol (100 µl of 100µM working solution in PBS) of fluorescently-labelled pHLIP or pHLIP variants on 1, 5, 10 dpi. Control mice were injected with PBS or unattached fluorophore. Two days after peptide injection, the mice were killed. Lungs, heart, spleen, liver and kidneys were harvested and fixed in 10% neutral-buffered formalin and then paraffin embedded for subsequent H&E staining or fluorescent imaging of tissue sections based on protocols described above. 2.2.20 Whole body and ex vivo whole organ bioimaging Mice were shaved with Veet hair removal cream and bioimaged on their ventral surfaces approximately every 24 h (± 2 h) post injection of fluorescently-labelled pHLIP for 8 days. Due to the presence of uneven colour pigment on C57Bl/6 mice skin, and the technical challenge in imaging lungs in a whole animal, organs were excised from pHLIP or pHLIP variant injected mice for ex vivo bioimaging. Excised organs were immediately placed in a petri dish or 24 well plates and imaged with 62 epifluorescence imaging function of IVIS spectrum (Perkin Elmer). The nearinfrared fluorophore Alexafluor-647 and Hylite fluor-647 was excited at 640nm and emission was analyzed at 700nm. Background signals from control mice injected with PBS were also analyzed. Region of interest (ROI) was drawn on the border of each organ and total radiant efficiency (units of [photons/sec/cm2/steradian]/[µW/cm2]) was quantified with Living Image software (Caliper Life Sciences). Signal-to-background radiant efficiencies were calculated by [(ROI 1- ROI 2)/ ROI 2] in which ROI 1 is the total radiant efficiency of the organ removed from peptide-injected mice, and ROI 2 is the total radiant efficiency of the organ removed from PBS injected mice (autofluorescence). To calculate contrast index (CI) with a normal (non-diseased) tissue, muscles on the right hind limb was excised and bioimaged to obtain a background fluorescence. Muscle fluorescence gives an indication of the amount of background fluorescence in tissue that should not accumulate peptide. CI is quantified as [(ROI 1 organ and ROI 1 muscle Organ - ROI 2 Organ)/ (RO1 Muscle - ROI 2 muscle)], in which ROI 1 are the total radiant efficiencies of individual organs (lungs, kidneys, liver, spleen etc) and muscle respectively excised from each peptideinjected mouse. ROI 2 organ and ROI 2 muscleare organ and tissue autofluorescence of PBS control mice (no peptide). CI provides an indication of the relative strength of the fluorescence signal in each organ. 2.2.21 Feature extraction and pHLIP quantification Infiltrated regions were identified by setting a threshold to segment darker stained regions of infiltrated regions from uninfiltrated regions based on images of H&E 63 stained whole lung sections (n = 2 sections per mouse). Within the infiltrated region, a smaller threshold was then applied to demarcate heavily and moderately infiltrated regions. Masks of heavily infiltrated, moderately infiltrated, and uninfiltrated regions were generated. Details of the quantification algorithm, designed based on infiltration index calculation, are described in “Lung histology, infiltration index calculation and pathology analysis” (subsection 2.2.3). Moderately infiltrated regions were confirmed at 20 - 40x magnification by thickening of alveolar walls, and presence of infiltration in the alveolar spaces that involves less than 70% of the total lung parenchyma. These regions were visibly distinguishable from heavily infiltrated sites, which are defined as densely stained regions with infiltration occupying more than 70% of the lung parenchyma. Extent of colocalization of Alexafluor 647-conjugated pHLIP with uninfiltrated, moderately and heavily infiltrated regions was assessed. Total pHLIP was identified by thresholding the intensity of each pixel in Alexafluor 647-conjugated pHLIP containing images. Percentage of pHLIP in heavily infiltrated region, moderately infiltrated region and uninfiltrated regions are calculated by the ratio of pHLIP within each mask to total pHLIP respectively. AEI cells were identified by thresholding the intensity of each pixel in tissue sections stained with anti-Pdpn. Holes among AEI cells, representing the alveolar spaces, were filled to form the mask of healthy areas. Damaged areas depleted of Pdpn positive AEI were segmented by subtracting healthy areas from total lung area (DAPI mask) in each section. The percentage of pHLIP-positive pixels within healthy and damaged regions was computed. Two whole lung sections from each mouse (n = 3 mice for uninfected or PR8-infected groups) were used for calculation. 64 Similarly, Prdx6 positive regions were identified by thresholding Prdx6 intensity. The percentage of total pHLIP-induced pixels was calculated in Prdx6 positive regions and Prdx6 negative regions. All image processing and computation algorithms were performed with Matlab (The Math Works, Inc., Natick Massachusetts). The codes for the algorithms are available on request. 2.2.22 Statistical analysis Quantification data were analyzed with Mann-Whitney U test and western blot analyses were performed with Wilcoxon signed ranked test using Graphpad prism unless otherwise stated in the figure legends. Data are represented with individual data points and median, Mean± SEM or Mean (SD) as stated in figure legends. Statistic differences are represented as *p [...]... results 2 provide new insights into the molecular pathology of influenza pneumonia, and offer opportunities to improve the management of influenza- induced lung disease 3 List of figures Figure 1.1 Structure of an influenza virus 10 Figure 1.2 The life cycle of influenza virus 11 Figure 1.3 Inflammation -induced ROS / RNS damages cellular molecules 25 Figure 1.4 Structures of modified nucleobases... histones that form DNA repair foci at the sites of DNA strand breaks in the lung tissue of infected mice DNA repair foci are induced in lung epithelial cells and infiltrating immune cells for prolonged duration even after viral clearance, suggesting that DNA damage is driven by inflammation Notably, DNA damage was observed in lung tissue especially during the tissue regenerative phase of infection,... Investigation of DNA damage as a potential mechanism of influenza pathogenesis 66 3.1.2 Phosphorylation of H2AX (Ser-139) during DNA strand breaks 69 3.1.3 DNA damage is triggered by influenza infection in vitro .72 3.1.4 Aims of study 74 3.2 Results 76 3.2.1 Characterization of H1N1 murine model 76 3.2.2 Prolonged inflammation after suppression of virus load... also evaluated One of the known mechanisms for ROS / RNS -induced toxicity is the ability to induce DNA lesions, which can lead to mutations and cell death In the first study, we investigated whether DNA damage, a potential consequence of inflammationinduced ROS / RNS, is associated with severe influenza pneumonia Using immunofluorescence techniques, we observed an increase in DNA damage, revealed by... viruses [e.g Influenza A/Puerto Rico/8/37 (H1N1) and Influenza B/Yamagata/16/88] (WHO 1980) Figure 1.1 Structure of an influenza virus Influenza is an enveloped virus composed of two major surface glycoproteins, haemagglutinin (HA) and neuraminidase (NA), and a minor component M2 The genome of influenza consists of 8 RNA segments, which are folded into ribonucleoprotein complexes (RNP) and encode for nucleoprotein... approved for adult treatment by Health Science Authority Effective only for influenza A, but not influenza B High resistance (> 99%) in recent circulating H3N2 and 2009 pandemic H1N1 Effective only for influenza A, but not influenza B High resistance (> 99%) in recent circulating H3N2 and 2009 pandemic H1N1 Influenza Antiviral Drug Resistance (2015, January 8) Centers for Disease Control and Prevention.Retrieved... METHODS AND MATERIALS 42 2.1 Materials 42 2.1.1 2.1.2 2.1.3 2.1.4 2.1.5 2.1.6 2.2 Media, chemicals and reagents .42 Cell cultures .43 List of antibodies for western blot and immunofluorescence .44 List of antibodies for flow cytometry 45 Source of influenza viruses .45 pHLIP and pHLIP variant peptide sequences and conjugation .46 Methods and protocols... Symmetrel® and Flumadine®, have been in use for therapeutic purposes for a longer duration of time, but are no longer recommended as prescriptions for influenza therapy and prophylaxis This is because amantadine and rimantadine, the compound names for Symmetrel® and Flumadine® respectively, are found to be ineffective during recent outbreaks with a widespread emergence of resistant H3N2 and 2009 pandemic... transmissibility, pathogenicity and infectivity in hosts Many subtypes of Influenza A viruses exist, and they are classified based on their expression of different HA and NA antigens Up to early 2015, a total of 18 HA and 11 NA antigens have been identified among influenza A viruses isolated from humans, birds and bats (Freidl et al 2015) Based on the expression of different HA and NA antigens, Influenza A viruses... characterization of pulmonary inflammation .81 3.2.4 Oxidative stress is elevated during infection 84 3.2.5 Host responses induce DNA damage in lung epithelium after influenza infection 86 3.2.6 DNA damage occurs in immune cell populations 94 x 3.2.7 Influenza infection induces polymerization of poly (ADP- ribose) 96 3.2.8 Influenza infection leads to apoptosis 99 3.2.9 Infection increases DNA damage